Using Subepidermal Moisture Level as an Indicator of Early Pressure Damage to Local Skin and Tissue

Abstract

OBJECTIVE 

Inflammation triggered by tissue injury increases blood flow to the injured area, resulting in fluid accumulation below the epidermis known as subepidermal moisture (SEM). This accumulation can be assessed using an SEM scanner. The primary aim of this study was to determine whether levels of SEM from repeated measures at a localized area confirm the absence of a pressure injury at that site in healthy participants.

METHODS 

Fifty patients attending a specialized pain clinic who were at very low risk of or unaffected by pressure injuries were enrolled. Subepidermal moisture readings were gathered at the center and near-proximate to the bony prominences of the heel and sacrum.

MAIN RESULTS 

Among the participants, repeated localized measurements did not differ significantly at or around a single anatomic site. The heel had a slightly higher variation than the sacrum, although the variance was less than 0.6 SEM units, indicating that there is likely no inflammation at these sites in patients who are not at risk of pressure injuries and who show no visual signs of localized tissue inflammation. Demographic and other potential confounder data were gathered, but no statistically significant findings emerged from these data.

CONCLUSIONS 

The relatively flat pattern of SEM values around the bony prominences of the heels and sacrum in healthy participants supports the hypothesis that healthy tissue is not inflamed.

INTRODUCTION

A pressure injury (PI) can cause pain and distress and increase treatment time for patients. In England, between 1,700 and 2,000 patients per month develop a PI, with an estimated associated cost to the National Health Service of more than £3.8 million per day.¹ In the US, the annual cost of PI treatment is an estimated $26.8 billion, and the average cost of a PI (per hospital admission) is $43,000.²

Because PI incidence serves as a quality indicator for healthcare facilities,³ patients are screened on admission for PI risk. At-risk patients are then assessed by healthcare practitioners using visual and palpation tests.³ The patient’s PI risk is usually determined by the use of risk assessment tools/scales such as the Braden⁴ or Waterlow,⁵ although these are subjective measures.^6–8 However, by the time the injury is visible, tissue damage has already taken place;⁹ perpendicular and shear forces contribute to tissue deformation and damage.¹⁰ The resulting breakdown of blood vessels and tissue cells causes an acute inflammatory reaction that increases both vasodilation and the permeability of blood vessels.¹¹ Plasma leaks into the interstitial space creating localized edema, or subepidermal moisture (SEM).^7,11

Given that tissue damage occurs before it is visible, the ability to determine the presence or absence of PI when it occurs would be useful. Oliveira et al¹² undertook a systematic review of the literature pertaining to ultrasound, thermography, and SEM for predicting PI. The four SEM studies included indicated that SEM values increased with tissue damage and that the sacrum and the heels were the most common anatomical locations for the development of erythema and stage 1 PI. The authors concluded that SEM is a promising method for the detection and prediction of early tissue damage and PI, although further studies are required.

The SEM Scanner 200 (Bruin Biometrics LLC, Los Angeles, California) is a medical device that aids in the identification of pressure damage by measuring SEM. The device measures tissue dielectric properties from an electrode placed on the skin, which records changes in biocapacitance. These changes are related to changes in the volume of fluid in the skin and tissues under the area being assessed.¹³ Recorded scanner absolute values range from 0.3 to 3.9 units, where higher values indicate increased SEM and lower values, a deficit. In ordinary clinical use, repeated measurements are made at the site of suspected injury and contiguous locations to gauge the relative difference in SEM. The difference between the highest and lowest readings, the delta (Δ), is used to assess PI risk at measured anatomies. A Δ below 0.6 is indicative of low site-specific SEM variance, or healthy tissue. A Δ equal to or greater than 0.6 reflects increased site-specific SEM variance and is indicative of developing pressure damage.

Separate studies have demonstrated the effectiveness of the SEM biomarker and the SEM scanner in the detection of potential PI. An elevated SEM level has been found to indicate tissue damage 3 to 10 days before visual skin damage or PI formation^7,14–17 and thus may serve as an early indicator.^7,12,14–17 Smith¹⁸ evaluated the SEM 200 scanner for 2 months with patients deemed at risk of PI (Waterlow score over 10). The heels and sacrum were scanned once a day from admission; patients were deemed to have early pressure damage if the SEM Δ was 0.6 or higher. All 35 patients had a Δ 0.6 or higher, and all received preventive interventions. Although none went on to develop new PIs during their inpatient stay, one developed a PI within 7 days of discharge. The author concluded that the SEM scanner can inform clinicians about risk before the visual signs of deterioration occur and help direct care staff mitigate PI risk. Using the SEM scanner in a PI prevention pathway in a Canadian general hospital, Raizman et al¹⁹ demonstrated a reduction in hospital-acquired PI (HAPI) incidence by 90% compared with baseline. Littlefield and Kellett²⁰ integrated Waterlow risk scores and SEM deltas in patient PI assessments at admission, during stays, and at discharge at two community hospitals in the UK. Results of their SEM-aided PI prevention and management protocol showed that stages 2 through 4 HAPIs were reduced by more than 95%; 75% of nurses described the scanner as easy to use, and 88% of nurses stated it provided additional information to support clinical decision-making. Finally, Okonkwo and colleagues’²¹ study involving 189 at-risk patients from 10 clinical study sites across the UK and US demonstrated that SEM measurements can accurately detect incipient pressure damage and that 87.5% of patients who developed PIs were previously detected by SEM. Importantly, SEM deltas identified early signs of pressure damage under the skin 5 days (median) earlier than visual skin assessment.²¹

To build on this previous research, this study’s primary objective was to collect SEM readings from healthy tissue (as determined by visual assessment and palpation tests by a qualified practitioner) both at and contiguous to the bony prominences of the sacrum and heel of patients without PIs. The study endpoint was to verify that healthy skin and tissue have a generally constant level of SEM (consistent absolute values and a Δ < 0.6) at the assessed anatomy, indicating the absence of localized inflammation. This is an important comparator for patients who are at risk of PIs or deep-tissue PIs, and where SEM deltas are elevated (Δ ≥ 0.6) but where the assessed skin and tissue are not exhibiting color, temperature, and/or edematous changes associated with PIs.

METHODS

A total of 50 patients from a private physician’s office in Virginia Beach, Virginia, who were not at risk of PI development were enrolled in this study. Recruitment was performed by the principal investigator. Patients were older than 55 years, a population considered at risk of reduced mobility,²² a risk factor for PIs in adults.²³ (The investigators also sought to compare results of this study to other, related studies involving patients with similar demographic profiles.) An equal number of males and females were enrolled. Exclusion criteria included broken skin at the anatomical locations being assessed, rheumatoid arthritis, gout, an autoimmune disorder, and use of corticosteroids.

The study conformed to the Federal Code of Regulations Part 812²⁴ and ISO 14155.²⁵ Ethics approval was obtained for the study (Quorum Review Inc, Seattle, Washington; QR#29154/1), and informed consent from each patient obtained. No follow-up procedures were required in the study, and no participants withdrew from the study. Consequently, no replacements of participants for analysis were made. A nominal stipend was provided to enrollees upon completion of study activities.

Data Collection

All assessments were performed during a single office visit lasting less than 2 hours. Consent, screening, and enrollment occurred on the same day as or immediately prior to the assessments. All data were recorded on case report forms (CRFs), and CRFs and datasets were audited by a site monitor. Minor findings by the site monitor were corrected in real time. A post hoc audit by a regulatory agency revealed no additional audit findings.

Demographic data (date of birth, sex, race and ethnicity, physical characteristics [height and weight], and medical history including comorbidities and medications considered potential confounders of SEM readings) were collected for each participant at the time of assessment. Prior to SEM scanner readings, a skin and tissue assessment consistent with those recommended in clinical guidelines^23,26 was performed. The results and associated diagnosis (no PI) was recorded on the CRFs.

The SEM scanner readings were obtained by the principal investigator who had received previous training on the use of the device and was blinded to the interpretation of SEM values. Patients were placed in a supine position and then repositioned for subsequent readings. No wait/acclimation time was required before SEM readings were taken, and readings were taken in immediate succession.

Readings of absolute SEM values were collected from the center of and contiguous to the bony prominences of the heel and sacrum. Each reading captured in the CRFs for each box (Figures 1 and 2; sacrum and heels, respectively) was the average of three measurements. Readings for the sacrum were collected at the sacral spine S4 (Figure 1, box C), at 4 cm (Figure 1, boxes 5, 6, and 8) and 8 cm (Figure 1, boxes 9, 10, and 12) away from the center of the bony prominence in each of the proximal, right lateral, and left lateral directions. For the heel, readings were collected at the calcaneus (Figure 2, box C) and at 4 cm away from the center of the bony prominence in the plantar, medial, and lateral directions (Figure 2, boxes 9, 10, and 12). Readings over the Achilles tendon, being too far superior from the heel and lacking the fleshy tissue associated with PIs, were not obtained.

Figure 1.: SPATIAL GUIDE FOR SEM SCANNER 200 READINGS, SACRUMSource: Bruin Biometrics LLC, study protocol SEM200-004. Permission to reprint provided by Bruin Biometrics, LLC.

Figure 2.: SPATIAL GUIDE FOR SEM SCANNER 200 READINGS, HEELSource: Bruin Biometrics LLC, study protocol SEM200-004. Permission to reprint provided by Bruin Biometrics, LLC.

Statistical Analysis

Following site monitoring, all data were locked and entered into SAS 9.4 (SAS Institute, Cary, North Carolina) for analysis. The data from the heel and sacrum were considered and computed separately. A repeated-measures analysis of variance, suitable for evaluating within-participant variability, was used to compare SEM absolute values from the anatomies assessed, as well as from the centers of the bony prominence at the sacrum and heel to readings taken within the surrounding areas at the specified distances from the center. The influence of potential confounding factors (body mass index, race, and sex) was explored using repeated-measures analysis of variance. All statistical comparisons were made at the two-sided 5% level of significance. The exploratory nature of this study did not require adjustments for multiplicity. No interim analyses or subgroup analyses were planned or conducted.

RESULTS

A total of 50 participants were enrolled. Table 1 summarizes the characteristics of study participants by sex.

Table 1.: PARTICIPANT DEMOGRAPHICS

To determine the distribution and variation of SEM measurements among participants, comparisons were made at each point within a single anatomic location. Absolute values for each anatomic location are shown in Tables 2 and 3, respectively. Absolute SEM mean and median measures at the sacrum varied by no more than 0.2 SEM units within each measurement location (2.5–2.3, all participants) and no more than 0.4 SEM units (2.6–2.2, male participants) between measurement locations.

Table 2.: SEM SCANNER 200 READINGS, SACRUM

Table 3.: SEM SCANNER 200 READINGS, HEEL

Readings at and around the heel (Table 3) were generally lower than sacrum readings. The sites around the heel locations had more variability among readings, although the variance among mean (2.0–1.7, all participants) and median values (1.9–1.7, all participants) did not exceed 0.5 SEM units for either sex (2.0–1.5, male participants; 2.1–1.8, female participants). Readings on the medial side of the heel appeared lower than readings on the lateral side of the heel, although this difference was not statistically significant.

Figure 3 presents a box plot of the left (western) to right (eastern) readings at and around the sacrum; Figure 4 presents a box plot of the lateral to medial readings at the heel. Absolute values show a relatively “flat” spatial distribution of readings.

Figure 3.: DISTRIBUTION OF SEM SCANNER 200 READINGS AT AND AROUND SACRUMThe diamond in the center of the box represents the mean, the line in the center of the box represents the median, the top and bottom of the box are the 25th and 75th percentiles of the observed values, the upper and lower fences represent the most extreme point outside of 1.5 times the interquartile range, and the circles represent potential outliers.

Figure 4.: DISTRIBUTION OF SEM SCANNER 200 READINGS AT AND AROUND HEELThe diamond in the center of the box represents the mean, the line in the center of the box represents the median, the top and bottom of the box are the 25th and 75th percentiles of the observed values, the upper and lower fences represent the most extreme point outside of 1.5 times the interquartile range, and the circles represent potential outliners.

No adverse or serious adverse events were reported in the study.

Potential Confounders

A summary of the readings and sex comparison is outlined in Tables 2 and 3. Based on the Mann-Whitney-Wilcoxon P values, all measured anatomic locations were similar by sex except for the center of the bony prominence of the heel (U = 739.5, P = .0482, Table 3). Specifically, the reading taken at the calcaneus was lower in males than in females. However, no associations between sex and readings at the heel were observed.

Participant characteristics, medical history, and visual evaluation data were analyzed as potential confounders of SEM readings (Table 4). There were no statistically significant variations based on sex, age, race, ethnicity, body mass index, smoking status, diuretics, surgery, or osteoarthritis among participant measurements at the sacrum, and of those factors, only race indicated a variation between groups for heel measurements. However, calluses on the heel and race may be two potential confounders for heel readings (F test of the null hypothesis of no between-participant variability because of calluses, F = 15.87, P = .0002; F test of the null hypothesis of no between-subject variability because of race, F = 9.83, P = .003). These results demonstrate that the presence of calluses on the heel decreases SEM readings at the center as well as lateral and medial to the bony prominence, but not on the plantar surface.

Table 4.: SUMMARY OF CONFOUNDER EVALUATION

The results in this study population also suggest that readings may be lower in the heel for African Americans than for whites. To further explore this result, average measurements around the heel were compared by F and P values. Except for the lateral site location, statistically significant differences were noted, although the number of African American participants was low (n = 9).

DISCUSSION

The consistency of SEM variation observed around each anatomic location in this study among healthy participants not at risk of PI suggests an absence of anatomy-specific inflammation and a relatively even spatial distribution of SEM.

The SEM scanner employed in this study uses the delta between the highest and lowest measurements around an anatomic site to assess the tissue status because absolute SEM measurements between individuals can differ substantially. In this study, readings at and around the sacrum were similar and had a generally even, unfluctuating pattern, whereas readings at and around the heel were relatively lower in value and showed more variability. This difference may be attributed to asymmetry in the vasculature of the medial and lateral aspects of the heel²⁷ or to an as-yet unidentified confounder that might be common in persons older than 55 years.

Two potential confounders of readings were identified—calluses and race—both of which impacted heel but not sacral readings. Although the association between heel calluses and lower readings may have a biologic explanation given the thickened skin, the association between African American race and lower SEM readings is difficult to determine. It could be attributed to an as-yet unidentified confounder, such as undiagnosed peripheral vascular disease or diabetes.^28,29 Of note, whereas the presence of a callus appears to lower all but the plantar SEM reading, African American race lowers all SEM readings almost equivalently and does not change the observed spatial pattern of relationships. However, both associations are based on a limited number of participants, and one or the other result may be attributable to chance alone. Other potential confounders did not significantly impact SEM readings.

As previously shown, studies using an SEM scanner have demonstrated early indications of PI.^7,12,14–17 More recently, Lawrence and Hancock³⁰ presented real-world evidence that explored the use of the SEM 200 Scanner as a component of a PI prevention program. In 11 centers in Canada, the UK, and Spain, more than 35,000 patient data values were obtained from 731 patients. Results demonstrated that 95% of patients showed an SEM delta over 0.5 during the usage period at some time point or anatomic location; 56% of facilities had no HAPI during the SEM usage period, and 73% of the facilities saw reductions in HAPI of more than 85% when SEM monitoring was implemented; and on average, across all 11 facilities, the HAPI was 1.2%, representing a 77% reduction over study start.

Implications for Practice

The SEM device is a handheld wireless device that is easy to use; it has high reported interoperator agreement and reliability.³¹ It can be used in any care setting and by any healthcare professional trained in its use, and it requires only a simple cleaning and disinfection between patients.³² Over time, the battery may require replacing, which is done by the manufacturer free of charge. All of this means the device is relatively simple and cost-effective to introduce into clinical practice.

In the UK, the cost per patient scanned using this device for an average episode of care (inpatient length of stay) is £1.54. The National Institute for Health and Care Excellence suggests that although this may represent a small additional cost compared with standard care, using the scanner to detect pressure damage earlier and implementing targeted interventions based on the results will ameliorate the cost of PI treatment.³²

CONCLUSIONS

These data support the use of spatial (as opposed to single-point) readings using the SEM scanner at anatomic locations susceptible to PIs for the assessment of tissue status. These data further suggest that measures of SEM can be obtained from diverse populations for the assessment of skin and tissue status. The relatively “flat”/consistent spatial pattern of SEM absolute values found around the heels and sacra of healthy participants in this study supports the hypothesis that healthy tissue is not inflamed.

Additional research to gather SEM readings from patients diagnosed with PIs is needed to create a fuller clinical picture of SEM values for anatomies with confirmed PIs. Such comparison would help practitioners objectively assess which of their patients’ anatomies are at higher or lower risk of developing PIs, potentially resulting in targeted anatomy-specific treatment decisions.