A Prospective, Observational Study to Assess the Use of Thermography to Predict Progression of Discolored Intact Skin to Necrosis Among Patients in Skilled Nursing Facilities

Skin temperature may help prospectively determine whether an area of skin discoloration will evolve into necrosis. A prospective, observational study was conducted in 7 skilled nursing facilities to determine if skin temperature measured using infrared thermography could predict the progression of discolored intact skin (blanchable erythema, Stage 1 pressure ulcer, or suspected deep tissue injury [sDTI]) to necrosis and to evaluate if nurses could effectively integrate thermography into the clinical setting.Patients residing in or presenting to the facility between October 2014 and August 2015 with a pressure-related area of discolored skin determined to be blanchable erythema, a Stage 1 pressure ulcer, or sDTI and anticipated length of stay >6 days were assessed at initial presentation of the discolored area and after 7 and 14 days by facility nurses trained on camera operation and study protocol. Variables included patient demographic and clinical data, data related to the discolored area (eg, size, date of initial discovery), and temperature and appearance differences between discolored and adjacent intact skin. Skin temperatures at the discolored and adjacent areas were measured during the initial assessment. All facility pressure ulcer prevention and treatment protocols derived from evidence-based clinical practice guidelines remained in use during the study time period. Participating nurses completed a 2-part, pencil/paper survey to examine the feasibility of incorporating thermography for skin assessment into practice. Data analyses were performed using descriptive statistics (frequency analyses) and bivariate analysis (t-tests and chi-squared tests); logistic regression was used to assess associations among patient and pressure ulcer variables. Of the 67 patients studied, the overall mean age was 85 years (SD 10); 52 were women; 63 were Caucasian; and the top 3 diagnoses, accounting for 60% of the study sample, included neurologic (ie, cardiovascular accident/dementia [14, 21%]), cardiac-related (14, 21%), and orthopedic (13, 19%) conditions. Twenty-eight (28) participants were long-term care patients, and 39 were admitted as short-stay patients. The most frequently reported location of discolored intact skin on presentation was the heel (27, 40%). The mean temperature at the site of the discolored skin was 33.6˚ C (SD 3) and at the adjacent skin was 33.5˚ C (SD 2.5). The mean size of the areas of discoloration was 11 cm² (SD 21). Capillary refill of the discolored area was absent on initial presentation in 49 patients (72%), and demarcation of the discolored borders was evident for 45 (66%). Of the 67 patients, 30 (45%) experienced complete resolution of the discolored area. At day 7, 8 (16%) of the remaining 50 patients in the sample exhibited skin necrosis and at day 14, a total of 12 patients of the remaining 37 (32%) exhibited skin necrosis. At day 7, skin necrosis was significantly associated with admission to a subacute unit (P = 0.01) and at day 14 to negative capillary refill at initial presentation (P = 0.02). Regardless of skin temperature, negative capillary refill at presentation was significantly associated with skin necrosis at day 7 (P = 0.04). A dichotomous variable was constructed to examine patients with cooler temperatures at the site as compared to their adjacent skin and persons with warmer skin temperatures at the center of the discolored skin for the presence of skin necrosis at both day 7 and day 14. In multivariate analysis, patients with cooler rather than warmer skin temperatures at the center of the discolored area as compared to the adjacent skin were more likely to develop necrosis by day 7 (OR 18.8; P = 0.05; CI: 104-342.44). Participating nurses were uncertain about the feasibility of integrating thermography into practice. Larger prospective studies with more heterogeneous samples are needed to determine the validity of skin temperature measurement as a predictor of skin necrosis and the utility of implementing thermography into clinical practice.

Pressure ulcers are a clinical concern across the care continuum. Despite advances in technology and the implementation of formalized prevention programs, pressure ulcer prevalence rates in skilled nursing facilities (SNFs) across the United States are reported to range from 4.1% to 32.3%.^1,2 According to the results of the National Nursing Home Survey,^3-5 1 in 10 nursing home residents in the US acquires a pressure ulcer and more than half of hospitalized patients with a primary or secondary diagnosis of a pressure ulcer when discharged are admitted to a long-term care facility. 

The Advancing Excellence in America’s Nursing Home campaign⁶ states nearly 1.5 million Americans reside in nursing homes. Given that pressure ulcers occur most frequently in patients over the age of 65 and that older adults are the fastest growing segment of the population in the nation, pressure ulcer rates in this at-risk population likely will continue to rise over the coming decades.^4,7 Moreover, the financial burden for pressure ulcer care in the United States tops $11 billion,⁸ which includes the increase in both hospital length of stay⁹ and the risk of mortality.^10,11 

Pressure-related areas of discolored intact skin are classified as blanchable erythema, a Stage 1 pressure ulcer, or suspected deep tissue injury (sDTI). The presence of blanchable erythema denotes an area with exposure to pressure with a high probability of resolving once the source of pressure is removed.¹² Although a Stage 1 pressure ulcer, defined by the National Pressure Ulcer Advisory Panel¹³ (NPUAP) as nonblanchable erythema of localized intact skin usually over a bony prominence, is considered a heralding sign of pressure ulcer risk, it also is characterized as a potentially reversible, localized area of injury. According to a current guideline,¹³ deep tissue injuries begin at the muscle-bone interface and occur as a result of tissue deformation due to a short period of intense pressure and/or as a result of tissue ischemia due to prolonged periods of immobility.^14,15 These injuries present as purple or maroon areas of discolored intact skin or a blood-filled blister from damage to underlying soft tissue damage due to pressure and/or shear.¹³ Deep tissue injury can create a serious clinical problem because it cannot be detected using current evidence-based clinical assessment techniques at an early stage and can rapidly deteriorate into a deep, full-thickness pressure ulcer even when pressure ulcer prevention strategies have been implemented.¹³ Thus, the degree of tissue destruction referred to as sDTI can only be currently determined ex post facto of pressure ulcer occurrence. 

The science supporting the pathophysiologic changes inherent to the evolution in sDTI continues to advance; however, the timeframe from the time of initial injury to the appearance of the sDTI has not been firmly established at this juncture. Expert opinion based on clinical experience and case studies in the area of sDTI evolution have postulated a 48-hour timeframe from injury to first appearance and from 7 to 10 days to a full-thickness wound.¹⁴ Using the principles of forensic science, Farid¹⁶ proposed a 7-day to 14-day window from the time of injury until the emergence of visible signs of injury and the progression to necrosis. 

The current level of evidence and available assessment techniques make it difficult for the bedside practitioner to expediently distinguish whether a pressure-related area of discolored intact skin will ultimately resolve or whether it potentially represents sDTI in evolution. Additionally, blanchable erythema, Stage 1 pressure ulcers, and emerging areas of sDTI are difficult to detect in individuals with darker skin tones because changes in skin color may not be evident and may be easily missed by clinicians.¹³ 

One potential method proposed to help prospectively determine whether an area of skin discoloration will evolve into necrosis has been the measurement of skin temperature. Physiologically, nonviable tissue is not perfused; thus, the area exhibits a cooler temperature due to lack of blood flow. Therefore, in areas of sDTI that progress to full-thickness ulceration and necrosis, skin temperature is hypothesized to be cooler due to the lack of perfusion.¹⁷ 

The use of thermography has been studied for decades for its efficacy in pressure ulcer detection; it was introduced in the 1970s by Verhonick et al.¹⁸ In a small preliminary methodological study¹⁸ of 3 subjects, these researchers outlined the application of thermography as a modality to assist with the objective quantification of pressure ulcer development by examining the relationship between pressure and skin temperature. Sprigle et al,¹⁹ using a repeated measure design among a study sample of 65 inpatients and outpatients in the rehabilitation setting, measured skin temperature using perfusion monitors and temperature strips placed on the skin to validate temperature differentials. These researchers found temperature differences (cooler or warmer) at the site of the erythematous skin center as compared to the adjacent normal skin could indicate impending changes in skin integrity. In an observational, retrospective study of 85 hospitalized patients, Farid et al¹⁷ studied skin temperature using noninvasive infrared thermography as an adjunct to current visual (skin color) and tactile assessment (capillary refill) in acute care patients exhibiting pressure-related areas of discolored intact skin. Results of this study showed areas with a cooler skin temperature at the site of the discoloration as compared to the adjacent normal tissue significantly predicted the progression of the discolored area to necrosis by day 7 (OR 31.8, P = 0.001, CI: 3.8-263.1). In addition, discolored areas that exhibited both cooler temperatures and negative capillary refill were found to be positively associated with skin necrosis by day 7 (P <0.0002). 

The purpose of this study was to determine if skin temperature, measured using infrared thermography, could predict the progression of pressure-related areas of discolored intact skin to necrosis in SNF patients. The following research questions were addressed in this study: 

1. What is the timeframe from initial presentation of a pressure-related area of discolored skin to necrosis in a sample of SNF patients?
2. Is there a relationship between skin temperature at the site of a pressure-related area of discolored intact skin and progression to skin necrosis in a sample of SNF patients?
3. What factors predict the progression of a pressure-related area of discolored intact skin to necrosis?
4. Can thermography be efficiently integrated into clinical use by practitioners in SNFs? 

A prospective, observational study was conducted to determine the relationship between skin temperature at the site of a pressure-related area of discolored intact skin and the progression to skin necrosis. The study methodology was derived from the methodology previously employed by Farid et al.¹⁷ Institutional Review Board approval was obtained from all participating SNFs before study initiation. 

Setting. This study was conducted in 7 SNFs in a northeastern state of the US, with a geographic distribution representing the north, mid, and southern sections of the state. All SNFs included had earned a 5-star quality rating on Nursing Home Compare and at least a 4-star rating attained for RN staff at the time of the study inception.²⁰ 

Inclusion/Exclusion criteria. Patients (long-term care and subacute care) residing in or admitted to any of the study sites were considered for study participation if the following inclusion criteria were met: observed pressure-related area of discolored intact skin determined to be blanchable erythema, a Stage 1 pressure ulcer, or sDTI at initial presentation; and 2) anticipated length of stay >6 days (shorter lengths of stay would preclude follow-up assessments on day 7 and day 14). Exclusion criteria were patients: 1) with pressure ulcers involving observed alterations in skin integrity, including intact blisters (this study was to determine the progression of discolored intact skin); 2) with ulcerations suspected to be from an etiology other than pressure (ie, neuropathic ulcers, venous ulcerations, arterial ulcerations, vasculitic lesions, and moisture-associated skin damage); 3) actively dying; or 4) with a history of a pressure ulcer or other tissue damage at the site of the current area of discolored intact skin (the composition of scar tissue may preclude an accurate temperature reading). 

Data collection procedures. Data were collected from each participating facility over a 10-month time period from October 2014 through August 2015. All data were collected by nursing employees at each of the SNFs who were responsible for pressure ulcer assessment within their institutions. The nurses were educated on the data collection procedures and use of the thermographic camera by the research team and the research study consultant via formalized training sessions conducted before the study began. All data were recorded on paper/pencil data collection records designed for this study by the research team. All data recorded were de-identified to protect confidentiality of the patients.

The nurse data collectors at the individual sites identified patients who presented with or developed a pressure-related area of discolored intact skin. If the inclusion/exclusion criteria were satisfied, the data collectors completed the initial patient assessment with follow-up assessments conducted on day 7 and day 14.

All patients received the usual standard of care during the study period. All facility pressure ulcer prevention and treatment protocols (derived from evidence-based clinical practice guidelines) remained in use during the study. Evidence of adherence to current pressure ulcer prevention was determined through documentation in the patient’s medical record, through observation at the time of discovery of the area of discolored intact skin, and at each follow-up assessment. 

Variables collected at initial assessment only. The following variables were collected only at the initial assessment of the pressure-related area of discolored skin: 1) date/time of initial discovery of discolored skin (if the patient had more than 1 area, only the largest area was recorded and used for analysis); 2) size (cm²), measured with a paper measuring device; 3) anatomic location; 4) skin temperature of the pressure-related area of discolored intact skin assessed using the infrared thermographic device; 5) skin temperature assessed using the thermographic device of the normal adjacent skin area; and 6) ambient room temperature (measured using thermographic device). In order to improve the visibility and identification of the area in the camera lens, adhesive stars were applied to the margins of the discolored skin (see Figure 1). The adjacent skin area was defined as the area within 5 cm to 10 cm of the outer margin of the area of discolored intact skin. 

Other data abstracted from the medical record at the time of the initial assessment included age, gender, race, admitting diagnosis, long-term care or subacute care admission, comorbid conditions (defined as any of the following: cardiovascular disease, diabetes mellitus, chronic pulmonary disease, peripheral vascular disease, end-stage renal disease), and Braden pressure ulcer risk assessment scale scores (on admission to the facility and at the time of the initial study assessment). Data regarding body temperature (measured using standard facility thermometer), body mass index, body weight, and serum albumin or prealbumin closest to the time of the discovery of the pressure-related area of discolored intact skin also were collected on initial assessment. 

Variables collected at all 3 assessments. The following variables were collected on initial assessment of the pressure-related area of discolored skin and repeated on day 7 and day 14: 1) presence or absence of demarcation; 2) color of the pressure-related area of intact skin; 3) presence or absence of capillary refill assessed using direct palpation of the area of discoloration skin and defined as positive (<3 seconds) or negative (>3 seconds or no visible blanching); and 4) evidence of the use of pressure ulcer prevention strategies as described above. 

The presence/absence of skin necrosis and pressure ulcer stage (if present) also were recorded on day 7 and day 14. 

Device. The thermographic device employed for this study was the Flir i^™ (Flir Systems, Boston, MA).²¹ This thermographic instrument is a hand-held, noninvasive, noncontact infrared detection device with a temperature sensitivity of <0.1˚ C at a distance of 2 feet. Inter-rater reliability of the use of the thermography camera was performed at the initiation of the study and periodically during the study period to confirm user accuracy and to validate that operational issues were not influencing the temperatures obtained. This device was chosen because it was previously employed in a clinical study¹⁷; however, the reliability and validity of this device have not been established in the context of a clinical setting to date.

Integration into practice. In order to determine if thermography could be successfully integrated into practice in the SNFs, a pencil/paper survey was designed and distributed to the 7 primary nurse users (1 from each of the participating facilities) during the final 3 months of data collection. This survey was comprised of 2 sections: the first section addressed 10 aspects related to operationalization of the specific thermography camera using the ratings satisfactory/unsatisfactory/unsure, and the second part of the survey included 3 open-ended questions addressing the feasibility of integrating thermography into the existing clinical setting. 

Sample size calculation. Sample size calculation was based on previous work by Farid et al.¹⁷ Assuming a sample with a ratio of warmer discolored areas to cooler discolored areas to be 2:1, a sample size of 60 was required to achieve statistical power.

Data analysis. Data were entered into SPSS statistical software package for Windows, Version 21 (IBM, Armonk, NY) and exported into Stata 14 (Stata Corp., College Station, TX). Data analysis was performed using SPSS for descriptive statistics and Stata 14 for bivariate analysis and multivariate analysis. Descriptive statistics including frequency distributions for all study variables were calculated. T-tests and chi-squared tests were conducted to detect differences in the study variables between patients with skin necrosis and those that remained necrosis-free. The presence of necrosis at both day 7 and day 14 was compared among patients with cooler pressure-related areas of discolored intact skin to patients with warmer pressure-related areas of discolored intact skin using chi-squared and t-tests. In order to conduct this comparison, a dichotomous variable was constructed to identify patients with lower temperatures at the site of discolored skin than the adjacent skin and persons with higher skin temperatures at the center of the discolored skin. Logistic regressions were used to capture associations among patient and pressure ulcer variables and the development of skin necrosis.

Patient sample description. Seventy-three (73) patients were entered into the study; 6 were eliminated from the sample due either to transfer out of the facility or patient death before the second assessment on day 7, yielding a sample size of 67 patients for analysis. The overall mean age was 85 years (SD 10). The sample was predominantly female (52, 78%) and Caucasian (63, 94%). The top 3 diagnoses, accounting for 60% of the study sample, included neurologic (ie, cardiovascular accident/dementia [14, 21%]), cardiac-related (14, 21%), and orthopedic (13, 19%) conditions. Of the 67 patients in this study, 28 (42%) were long-term care patients and 39 (58%) were admitted as short-stay patients (see Table 1). 

The most frequently reported location of discolored intact skin on presentation was the heel (27, 40%). The mean temperature at the site of the discolored skin was 33.6˚ C (SD 3) and at the adjacent skin was 33.5˚ C (SD 2.5). The mean size of the areas of discoloration was 11 cm² (SD 21). Capillary refill was absent on initial presentation of the area of discoloration in 49 patients (72%), and demarcation of the discolored borders was evident for 45 (66%) (see Table 2). 

The discolored area completely resolved in 30 patients by either day 7 (17, 25%) or day 14 (13, 19%). The 2 most common stages of pressure ulcer, reported at both day 7 and day 14, were sDTI (19 [38%] and 10 [27%], respectively) and unstageable (8 [16%] and 11 [30%], respectively) (see Table 2). The skin color most frequently reported on initial presentation was purple (29, 43%); of the 29 patients who presented with purple discoloration at the first assessment, 10 (34%) of these areas resolved, 6 (21%) by day 7 and 4 (14%) by day 14. By day 7, 12 (54%) of the initial purple discolored areas were staged as sDTI and 6 (27%) were staged as unstageable pressure ulcers. At day 14, 5 (26%) were staged as sDTI and 8 (42%) were staged as unstageable pressure ulcers. Of the remaining skin colors on presentation, red and deep red, respectively, were found to be the subsequent 2 colors most frequently reported (19 [28%] and 10 [15%]) in this sample. Of those with deep red discoloration on presentation, 2 (20%) had resolved by day 7 and 1 (10%) by day 14. The most frequently reported pressure ulcer stage at day 7 for patients with deep red discoloration on presentation was sDTI (3, 30%). For persons with red discoloration at presentation, 5 (26%) resolved by day 7 and 4 (21%) by day 14. The most frequently reported stage of pressure ulcer for those with red discoloration at day 7 was Stage 2 (3, 16%). 

The temperature of discolored skin as compared to the adjacent area was warmer in 41 (61%) and cooler in 26 (39%). The analysis of the dichotomous variable constructed to differentiate patients with cooler from those with warmer areas of discolored skin as compared to the adjacent skin showed centers categorized as cooler had a mean temperature of 31.2˚ C (SD 3.7); for those categorized as warmer, the mean temperature was 35.2˚ C. Thus, the mean temperature difference between patients with warmer temperatures versus cooler temperatures was 3.99˚ C. 

Time between initial presentation of a pressure-related area of discolored skin and development of necrosis. Patients that did and did not develop skin necrosis were compared on all study variables on day 7 and day 14 (see Table 3). By day 7, 17 discolored areas had resolved. Of the 50 patients remaining in the study sample, 8 (16%) had skin necrosis. At day 7, all 8 patients who developed skin necrosis were subacute or short-stay patients (χ² = 6.52; P = 0.016) and admitted to the SNF with an emerging area of discoloration, suggesting the areas of discoloration were already evolving at the time of admission to the SNF. By day 14, 13 additional pressure-related areas had resolved. Of the 37 patients remaining in the sample, 12 (32%) had skin necrosis. 

Capillary refill. A statistically significant difference was found at day 14 between patients who did and did not develop skin necrosis with regard to capillary refill. Overall, 49 of the study participants (72%) had negative capillary refill on initial presentation of the discolored area. All of the 12 patients that developed skin necrosis by day 14 had negative capillary refill evident on the initial presentation of the area of discoloration (P = 0.02). Statistically significant differences both at day 7 and day 14 also were found between the stage of pressure ulcer reported (χ² = 25.17, P = 0.000; χ²= 13.1, P = 0.002, respectively); patients with negative capillary refill at the initial presentation of the area of discoloration more frequently developed pressure ulcers categorized as sDTI or unstageable (see Tables 4 and 5). 

Relationship between skin temperature at the site of a pressure-related area of discolored intact skin and progression to skin necrosis. No statistically significant difference was found between cooler and warmer temperatures on presentation and the development of skin necrosis at day 7 (P = 0.93) or day 14 (P = 0.38) in univariate analysis. Additionally, no statistically significant differences were found in univariate analysis for any of the study variables between patients with cooler versus warmer skin temperatures on initial presentation (see Table 6). 

Predicting the progression from discolored intact skin to necrosis. Multivariate analysis of the study variables showed admitting diagnosis and cooler discolored intact skin temperatures were explanatory variables. Cooler skin temperature increased the odds of experiencing necrosis at day 7 at the 92% confidence level (odds ratio [OR] = 6.3; P = 0.08, CI: 0.83-48.00). When demographic (eg, age and admitting diagnosis) and environmental variables (eg, ambient room temperature) were added to the baseline model, the OR for development of skin necrosis substantially increased and was statistically significant at the 95% confidence level (OR = 18.8, P = 0.05, CI: 1.04-342.44) (see Table 7). Although a statistically significant Pearson correlation was found between capillary refill and necrosis at day 14, this association did not emerge in regression analysis (see Table 3). 

Integrating thermography into clinical use. Results of the survey distributed to the participating nurses found the camera was easy to operate and the infrared images easy to retrieve. The average amount of time needed to incorporate thermography into the clinical assessment of an area of discolored skin was 3 to 5 minutes. However, when asked if they believed thermography could be easily implemented into practice, 5 of the 7 participating nurses (more than 70%) were unsure or believed thermography could not be easily integrated into practice. Some uncertainties stemmed from specific clinical situations, such as the amount of time required for patient set-up to acquire the temperature reading and the need for more than 1 caregiver to position a patient depending on the anatomical location of the discolored area. 

The possible null effect on current pressure ulcer prevention strategies also was identified; adding skin temperature to the assessment of pressure-related areas of skin had the potential to have no impact on current pressure ulcer practices. Compliance with camera use by future users, the possibility of camera breakage and loss, as well as the overall cost of the modality also were cited as possible concerns. 

Overall, results of this multivariate analysis suggest a relationship between lower skin temperatures of discolored intact skin as compared to the adjacent skin and skin necrosis by day 7 but not day 14. 

Review of the relevant literature offers varied results on the use of skin temperature as an adjunct to present clinical assessment techniques to detect deep tissue injury and pressure ulcer development. In the study by Farid et al,¹⁷ patients assessed using thermography who exhibited cooler rather than warmer skin temperatures were almost 32 times more likely to experience skin necrosis. In 2 previous clinical studies,^19,22 temperature variations either cooler or warmer at the site of the discoloration as compared to the surrounding skin were found to be associated with pressure ulceration. Sprigle et al¹⁹ found a 1˚ C difference in temperature (either warmer or cooler) between an area categorized as a Stage 1 pressure ulcer and the adjacent skin could herald a change in skin integrity. Similarly, Judy et al²² compared the use of the Braden scale to temperature measurement obtained through the use of infrared technology to determine pressure ulcer risk in a small study of hospitalized patients (N = 5) using a repeated measures design and found that a variance in skin temperature of 1.5˚ C (either cooler or warmer) between the bony prominence and the adjacent skin was more accurate in predicting pressure ulcer development of the heels and the sacrum in patients at risk for pressure ulcers than use of the Braden Scale alone. 

Laboratory studies of the efficacy of skin temperature measurement have shown variations in skin temperature (either warmer or cooler) were associated with skin necrosis; in one,²³ the use of infrared thermography was determined by the researchers to be a useful modality for the objective measure of early sDTI of the heel. However, the researchers concluded the temperature of the sDTI could be normal, elevated, or decreased when compared to adjacent healthy tissue, with increases in temperature observed with inflammation and temperature decreases with tissue ischemia. The researchers concluded temperatures in a suspected area of DTI were initially cooler, but with reperfusion and the presence of a severe inflammatory response, the temperature could increase in an ischemic area. Thus, the time at which the temperature is taken can be crucial to the determination of evolving necrosis. The challenge this presents in the clinical arena is proactively identifying patients at high risk for sDTI, accurately timing the measurement of skin temperature (initial moment of presentation or in the reperfusion phase) at which to capture the temperature fluctuation that the researchers purported to occur, and translating this finding to bedside practitioners so as to be clinically relevant. 

Interestingly, in another laboratory study using laser Doppler flowmetry,²⁴ reactive hyperemia characterized by higher temperatures in combination with pressure were found to be potential contributors to skin ischemia. The researchers proposed that lowering skin temperature proactively may have the potential to prevent pressure ulcer development; however, this study was limited by the participation of healthy persons who did not have the same comorbid burden of illness characterized by many patients, especially SNF patients who are at high risk for pressure ulcer development. 

In the current study, the pressure ulcer stages sDTI and unstageable were the 2 most common at both day 7 and day 14, with purple discoloration reported as the most common skin color at initial presentation of the discolored area. Among patients with areas of purple discoloration of intact skin on initial presentation, 6 (21%) had resolved at day 7, 12 (41%) were staged as sDTI, and 6 (21%) were categorized as unstageable. In other studies examining the evolutionary timeframe of sDTI, results have been varied. In a prospective, exploratory study of 40 hospitalized patients, Richbourg et al²⁵ found that of the patients who presented with purple discoloration, 2 (5%) had reepithealized, 18 (43%) had no change in this color, and 7 (17%) had progressed to slough or eschar. However, time to follow-up in this study varied from 1 to 20 days (average of 6 days). Moreover, in a retrospective study of 77 hospitalized patients conducted over a 2-year period, Sullivan²⁶ found that at follow-up (average 6 days, range 1 day to 4 weeks) from initial presentation 85 (66.4%) of the areas of sDTI had resolved or were moving toward resolution, 31 (24.2%) remained unchanged in appearance, and 12 (9.3%) had evolved into full-thickness wounds. 

The divergent findings between the previous and current research could be a reflection of different methodologies used, or they may suggest that DTI evolution may be variable. Factors that affect the evolution of sDTI still need to be elucidated, and staging a sDTI in evolution can be challenging.¹⁴ Thus, further empirical investigations that will contribute to the body of knowledge regarding the underlying pathogenesis of sDTI are needed. 

In the current study, negative capillary refill on initial presentation was found to be significantly associated with necrosis at day 14 and more frequently associated with pressure ulcers staged sDTI and unstageable at both day 7 and day 14. Farid et al¹⁷ found negative capillary refill in combination with cooler temperatures to be strongly associated with skin necrosis at day 7 (P <0.0002). Current results also revealed 18 patients (100%) with positive capillary refill with either warmer or cooler skin temperatures did not develop skin necrosis by day 7. Moreover, all 8 patients who developed skin necrosis by day 7 had negative capillary refill at initial presentation, regardless of skin temperature (χ² = 8.2055, P = 0.04) (see Table 8). 

Capillary refill is used as a measurement of peripheral perfusion. Obstruction of capillary blood flow can result in tissue ischemia.¹² Once the pressure is relieved, reactive hyperemia can ensue, resulting in erythematous appearance of the skin. On palpation, this erythema can be blanchable, with probable resolution if the source of pressure is relieved. However, nonblanchable erythema and corresponding changes in skin color to dark red or purple has the potential to be an indicator of deeper tissue damage.¹² Consistent with this description, the current study found 39 of the patients (82%) who had negative capillary refill of the discolored area on initial presentation exhibited initial skin color changes described as deep red, purple, or brown, although these initial skin color changes were not statistically significantly associated with necrosis at either day 7 (P = 0.20) or day 14 (P = 0.15). Regardless of the presenting color of the skin or temperature, the current study found capillary refill was better able to detect the progression of a discolored area of skin to necrosis.

Based on the results of the survey conducted with the nurses regarding use of thermography in the SNFs, a degree of uncertainty exists as to whether thermography could realistically be operationalized in the SNF setting. In previously published studies,^17,19,22 thermography was primarily used as a research tool, with camera use limited to the primary study investigator and trained research assistants. This appears to be the first study to involve nurses trained on the use of the thermography cameras. Although the cameras were found to be easy to operate, participating clinicians suggested cameras should be small enough to fit into the pocket in order to expedite the assessment. The need for an additional caregiver to assist in positioning the patient for the thermographic reading was cited as a disadvantage, increasing the time required to obtain the temperature of the discolored area. Cost for each camera (approximately $2000.00) and the potential for camera breakage and loss also were cited as potential concerns for implementation. Thus, based on the feedback of these users and coupled with the results of this study, the ability to successfully integrate thermography as part of a routine physical assessment at this time remains questionable. As technology continues to advance, future research will be warranted to explore use of thermography cameras that have greater portability (ie, pocket-sized), are less costly, and provide for greater ease in patient positioning in order to fully evaluate the pragmatic use of this modality in the clinical setting. 

The sample size may be considered a limitation of this study; however, the minimum sample size required based on the calculation proposed by Farid et al¹⁷ was achieved. Also, the lack of darker-skinned patients in the sample could be considered a limitation because it remains unknown whether the use of thermography may be able to detect early skin changes that can lead to necrosis in patients in which changes in skin color and capillary refill are difficult to assess. Longer observation periods also may be warranted to detect further deterioration in discolored areas to necrosis. In addition, the type of infrared thermography device selected may have been a limitation; however, the authors employed the same device used in the study by Farid et al¹⁷ to remain consistent in methodology. The homogeneity of the sample in SNF setting presented a challenge to establishing associations between some of the independent variables and the outcome and thus limits the generalizability of the findings. All findings should be interpreted with this in mind. Further research is warranted using larger and more heterogeneous sampling. 

Cooler skin temperature at the site of the discoloration emerged as a significant predictor of the evolution of an area of discoloration to necrosis at day 7 of the 14-day study period. In addition, negative capillary refill, measured by detection of skin blanching, was significantly associated with skin necrosis by day 14. Notably, 100% of patients with positive capillary refill on presentation of the pressure-related areas of discoloration did not progress to skin necrosis at the discolored site, suggesting capillary refill is a reliable and simple technique that remains an important parameter when assessing pressure-related areas in patients with lighter skin tones. Assessment of capillary refill, currently part of the standard skin assessment for patients that present with areas of discolored skin, is a procedure that takes <1 minute to perform and can provide the clinician with real-time information regarding the potential viability of the skin. 

Infrared thermography continues to gain popularity but is still in its infancy with regard to its practicality and effectiveness as a screening tool for the development of necrosis in pressure ulcers in the real-world clinical setting. Moreover, the utility of successfully assimilating thermography into broad clinical practice remains largely unknown at this time because of a lack of an empirically established temperature differential between the adjacent skin and the area of skin discoloration that would guide clinicians in making the determination if any area of discoloration would progress to necrosis from those areas that may ultimately resolve. Larger studies across diverse clinical settings are warranted. 

Whether all pressure ulcers are avoidable or if the effects of pressure can be mitigated in certain clinical situations has been debated and scrutinized.^27,28 Continued empirical study into innovative, prospective, objective clinical assessment tools will aid in answering this question. If the clinical progression of a pressure-related area of discolored intact skin can be determined earlier in the course of ulcer evolution, earlier application of current pressure ulcer prevention strategies holds the possibility of improving patient outcomes, or conversely, earlier detection may add to the current knowledge base regarding the unavoidable pressure ulcer. The challenge remaining for researchers and clinicians is to identify proactive pressure ulcer assessment techniques that are reliable and valid, yet clinically feasible, and easily integrated into practice across diverse patient care settings. 

The authors acknowledge the following nurses for their assistance and diligence in the data collection process for this study: Edward Acquah Jr, Kathy Brady, Sirner Dhaliwal, Judy Dedio, Kristin Melbourne, Sarah Robinson, Evelyn Savarese, Theresa Smith, and Norman Taclob.

Funding for this study was provided by the Centers for Medicare and Medicaid Services. 

Dr. Cox is an assistant professor of nursing, Rutgers University School of Nursing, Newark, NJ; and an advanced practice wound, ostomy, continence nurse, Englewood Hospital and Medical Center, Englewood, NJ. Ms. Kaes is Director of Quality Improvement and Clinical Services, Health Care Association of New Jersey, Hamilton, NJ. At the time of this study, Mr. Martinez was the Institutional Research Specialist, Rutgers University, School of Nursing, Newark, NJ; he is now a Senior Institutional Research Analyst, Emory University, Atlanta, GA. Mr. Moles is President, TRANSITION HealthCare Consultants; and President, Nursing Home Expert Opinion Services, Monroe Township, NJ.

Please address correspondence to: Jill M. Cox, PhD, RN, APN-C, CWOCN, Rutgers University, 869 Rivervale Road, River Vale, NJ 07675; email: jillcox@sn.rutgers.edu.