Reliability of the skin blotting method when used on the elderly

Abstract

The aim of this study was to compare protein secretion on intact skin of extremities and verify the relationship between the marker proteins on abdominal skin and systemic factors using skin blotting. A cross‐sectional study was conducted among elderly patients aged 65 years and older (N = 73) at a long‐term medical facility in Japan. Skin blotting was performed on the right and left forearms, right and left lower legs, and abdomen. Pearson's correlations and Bland–Altman plots were utilised for comparing the protein secretion from the skin between the right or left forearms or lower legs. Multiple regression analysis was applied to determine the relationship between intensity levels of 3 proteins on the abdominal skin and the systemic factors. Bland–Altman plots demonstrated that there was no significant difference between right and left secretion levels on the forearms and lower legs among 3 proteins. Multiple regression analysis showed that age and antiplatelet use was positively associated with decreased collagen type IV and increased matrix metalloproteinase 2 levels, respectively. Our findings suggested that collecting samples from either the right or the left skin would be sufficient if skin properties between arms and legs are evaluated using skin blotting.

1. INTRODUCTION

Skin tears are traumatic wounds caused by shearing, friction, and/or blunt forces that result in the separation of skin layers.1 Skin tears occur more commonly on the extremities of the elderly, and they can be painful and ultimately reduce patient quality of life.2, 3 Elderly patients with existing skin tears are most likely to experience additional tears, referred to as a skin tear recurrence.2, 4, 5, 6 In society with a rapidly growing proportion of elderly people, it is important to predict which population is at high risk of skin tears through non‐invasive and simple daily skin assessment.

The skin of the elderly is predisposed to skin tears because of a series of changes that occurs during the ageing process.7, 8, 9 Skin senescence includes the structural alteration of dermal–epidermal junction, such as flattening the dermal papillae and thickening the epidermal basement membrane.10 These changes impair the resistance of skin to external forces and increase the risk of injuries. Therefore, the structural evaluation of skin and the analysis of structure‐related factors will contribute to the identification of patients at high risk of skin tears.

To date, several non‐invasive methodologies of skin assessment have been developed. Although confocal laser scanning microscopy11 and ultrasonography12, 13 provide high‐resolution images of the dermal and epidermal structure, they require expert technique for examination and are too expensive to use for daily assessment. Tape stripping is widely used for measurement of the accumulated factors in the corneal layer.14 However, the repeated detachment of adhesive tapes creates a possible risk of tearing in elderly skin. Minematsu et al. have recently developed a novel skin assessment method, “skin blotting.”15 The hydrated nitrocellulose membrane on the skin surface attracts cytokines such as tumour necrosis factor α (TNF‐α) and nerve growth factor β; extracellular enzymes such as matrix metalloproteinase 2 (MMP‐2); and extracellular matrices such as collagen type IV (COL4) and fibronectin from epidermis, dermis, and subcutaneous fat layers, non‐invasively.15, 16, 17

We previously conducted skin blotting on the forearms of elderly patients with or without skin tears and demonstrated the significantly lower secretion of COL4, a major matrix of the epidermal basal membrane,18, 19 and MMP‐2, a proteolytic enzyme that degrades COL4,20 and significantly higher TNF‐α, a pro‐inflammatory cytokine that regulates collagen remodelling21 in elderly patients with skin tear. Skin properties are determined by systemic factors such as age, medications, and nutrition and external local factors such as chronic exposure to ultraviolet and mechanical stimulations. The systemic factors can equally affect the whole skin, whereas the external local factors alter the local skin conditions, resulting in the different skin properties according to a part. Our previous study, in which we indirectly estimated the skin properties related to skin tear at the contralateral side of forearm with skin tear,16 suggested that the right and left side of the extremities might be similarly affected by the local stimulations. However, the difference between the right and left side of the extremities has not yet been compared. Furthermore, the condition of abdominal skin is thought to be stable and associated with systemic factors. The verification of these hypotheses in the skin blotting examinations of marker proteins is a fundamental step to establish the prediction methods for patients at a high risk of skin tear. The aim of this study was to compare the protein secretion on the intact skin of extremities using skin blotting and verify the relationship between the secretion levels of marker proteins at abdominal skin and systemic factors among elderly patients at a long‐term medical facility.

2. MATERIALS AND METHOD

2.1. Study population

This cross‐sectional study was conducted in a long‐term medical facility with 500 beds in Ishikawa Prefecture, Japan, between December 2014 and March 2015. This specialised facility for elderly patients requiring chronic medical care was selected because almost all patients received equal daily support based on their individual physical condition, which enables researchers to assess the aged skin conditions. Patients were included in this study if they were aged 65 years or older, and they or their proxies gave informed written consent to participate in this study. Patients who had evidence of cutaneous disease on the extremities and abdomen, such as eczema herpeticum and psoriasis, or who had a risk of being disseminated intravascular coagulation (DIC) before dying were excluded. Altered skin structure because of these symptoms may be a threat to the internal validity of measuring the skin properties by skin blotting.

2.2. Data collection

2.2.1. Demographic data and risk factors

The systemic factors associated with skin tears development, including age, gender, body mass index, paralysis, stiffness, medication (polypharmacy, corticosteroid use, anticoagulant use, and antiplatelet use), nutrient intake, and diseases, were collected from the medical records and nursing records.1, 2, 3, 5, 6, 16 There were 3 ways of nutrient intake among participants: (1) oral intake, (2) tube feeding through gastrostomy tube or nasal tube, and (3) intravenous feeding through central or peripheral catheter. If some patients used 2 or more ways of nutrient intake, the intake of the highest calories was selected as the independent variable. Diseases such as diabetes mellitus, hypertension, and renal failure were used as independent variables related to skin and blood vessel fragility development.22, 23

2.2.2. Sample collection by skin blotting

Skin blotting was performed on the day the patient did not receive hygiene care and ointment application because these skin care techniques can affect the protein collection in skin blotting. Skin blotting was conducted at the following 5 regions in the lateral position (Figure 1): right and left forearms (posterior midpoint between lateral epicondyle and radial styloid process), right and left lower legs (upper quartile of the lateral lower legs), and abdomen (midpoint between the umbilicus and the iliac crest). If participants displayed skin injuries on these regions, we performed skin blotting at the peripheral points 5 or more centimetres away from the injury. A piece (12 mm × 22 mm) of nitrocellulose membrane (Bio‐Rad, Hercules, California, USA) was moistened with 20 μL of normal saline and attached to the skin surface with gentle medical adhesive tape (Nitto Medical, Osaka, Japan) for 10 minutes. Then, the medical adhesive tape was removed from the patient's skin with minimal stimulus. The blotted membranes were stored at 4°C until staining.

Figure 1.

The 5 anatomical locations on which skin blotting was performed: 1 and 2: right and left forearms, respectively (posterior midpoint between lateral epicondyle and radial styloid process); 3, abdomen (midpoint between the umbilicus and the iliac crest); and 4 and 5 right and left lower legs, respectively (upper quartile of the lateral lower legs)

2.3. Immunostaining

COL4, MMP‐2, and TNF‐α were the target proteins in this study. The following procedure of immunostaining was performed by SNAP i.d. 2.0 system (Merck Millipore, Billerica, Massachusetts, USA). The membranes were hydrated with phosphate‐buffered saline. To inactivate endogenous peroxidases, the membranes were incubated with 20% methanol supplemented with 0.3% hydrogen peroxide at room temperature for 10 minutes. Following inactivation of endogenous alkaline phosphatase and blocking with blocking buffer (Nacalai Tesque, Kyoto, Japan) supplemented with 500 μg/mL of levamisole hydrochloride (Sigma‐Aldrich, St. Louis, Missouri, USA) at room temperature for 10 minutes, membranes were reacted with primary antibodies and secondary antibodies for 10 minutes each. The primary antibodies for COL4 (Merck Millipore), MMP‐2 (Thermo Fisher Scientific, Waltham, MA, USA), and TNF‐α (Santa Cruz Biotechnology, Dallas, TX, USA) were diluted with blocking buffer at 1:250, 1000, and 200, respectively. The dilutions of both secondary antibodies, including alkaline phosphatase‐conjugated anti‐goat IgG (Jackson ImmunoResearch, West Grove, Pennsylvania, USA) and horseradish peroxidase‐conjugated anti‐rabbit IgG (Jackson ImmunoResearch), were 1:1000.

Immunoreactivities were visualised by incubating with chemiluminescent substrates for alkaline phosphatase (DuoLux, Vector Laboratories, Burlingame, California, USA) for 5 minutes and peroxidase (Luminata Forte Western HRP substrate, Millipore) for 2 minutes and were captured by LumiCube (Liponics, Tokyo, Japan). The exposure times were 30 seconds for TNF‐α and MMP‐2 and 3 seconds for COL4. Simultaneously, bright field images of membranes were also recorded.

2.4. Image analysis

The colour images were separated into RGB channels using Image J (National Institutes of Health, Bethesda, Maryland, USA). The blue channel of each of the 3 proteins was selected for image processing. The measurement area was identified by overlaying the corresponding bright field images by GNU Image Manipulation Program 2.6.5. Finally, mean brightness of the luminescence was measured using ImageJ.

2.5. Ethical considerations

This study was approved by the Ethical Committee of the Graduate School of Medicine, University of Tokyo, Tokyo, Japan, and conformed to the ethical guidelines of the 1975 Declaration of Helsinki. Proxies provided signed informed consent. The consent form included a description of the project; the procedures involved; and the risks, benefits, and costs to the participant. All measurements were performed by the primary researcher, a fully trained research nurse from the Department of Wound Care Management. Initially, the primary researcher made every effort to establish an intimate rapport with all elderly patients with dementia because they may often develop a new injury because of rejection or fear of strangers (primary researcher). The primary researcher kept an eye on some elderly patients who were predisposed to removing the tape by themselves when attaching the membranes (10 minutes).

2.6. Statistical analysis

Each set of values represent the median and top 25th percentile (first quartiles) and bottom 25th percentile (third quartiles). The statistical significance level was set at P = .050. The statistical analyses were performed using the JMP for Windows ver. 10.0.2 software programme (SAS Institute, Japan) and SPSS 15.0.1 for Windows (SPSS Inc., Chicago, Illinois).

The bilateral comparison of skin blotting measurements on the forearms and lower legs were assessed using Pearson's correlation coefficient and the Bland–Altman plots with 95% limits of agreement.24 Bland–Altman analysis plots the difference between the 2 measured values on the y‐axis (difference, a − b) and the average value of the measured values on the x‐axis ((a + b)/2).

Multiple regression models were applied to determine the relationship between the intensity of 3 proteins on the abdominal skin and the systemic factors. Abdomen is a part that experiences minimal impact of external factors. Independent variables that were correlated with outcome (P < .200) by Pearson's correlation coefficient were included in the multiple regression analysis as candidate factors. If strong correlations between independent variables were found, 1 of the variables was entered into the model to avoid multicollinearity. The fitness of the model was assessed using adjusted R ².

3. RESULTS

3.1. Study participants

Seventy‐three patients were recruited. Among them, 1 patient with serious pneumonia and 2 who did not provide consent because of pain were excluded from analysis. Therefore, 70 patients comprised the participant pool for this study.

The median age (min and max) of patients was 86.5 (81‐90) years, and 64.3% of the participants were women. Regarding diagnosis, stroke and dementia were noted in 88.6% and 25.7%, respectively. Hypertension, diabetes mellitus, and renal failure related to skin and blood vessel fragility were noted in 14.3%, 11.4%, and 5.7%, respectively (Table 1).

Table 1.

Characteristics of participants

		(n = 70)
Age (y)	86.5	(81‐90)
Gender
Male	25	(35.7)
Female	45	(64.3)
Body mass index
~18.5	48	(68.6)
18.6~	22	(31.4)
Paralysis	4	(5.8)
Contracture of the arm	51	(72.9)
Contracture of the leg	53	(75.7)
Medications
Steroid use	4	(5.7)
Anticoagulant use	3	(4.3)
Antiplatelet use	13	(18.6)
Numbers of medicine	4	(0‐18)
Nutrient intake
Tube feeding	27	(38.6)
Intravenous feeding	22	(31.4)
Oral intake	21	(30)
Diseases
Stroke	62	(88.6)
Dementia	18	(25.7)
Hypertension	10	(14.3)
Diabetes mellitus	8	(11.4)
Renal failure	4	(5.7)

Values are median (top 25th percentile‐bottom 25th percentile) or number of patients (%).

3.2. Secretion level of 3 marker proteins

A boxplot shows the detection levels of COL4, MMP‐2, and TNF‐α at 5 sampling sites (Figure 2). The median intensity levels (first and third quartiles) of 3 proteins at the right and left forearm, abdomen, and right and left lower legs were as follows: COL4: 8.8 (7.1‐62.4), 8.6 (7.2‐36.8), 8.3 (7.0‐17.1), 8.5 (6.7‐51.0), and 8.0 (6.4‐57.5); MMP‐2: 12.2 (9.1‐110.9), 12.6 (8.7‐161.5), 12.4 (7.8‐77.2), 12.3 (8.4‐99.1), and 11.9 (8.1‐185.5); and TNF‐α: 13.2 (8.2‐39.8), 13.3 (9.5‐26.9), 12.4 (8.7‐72.6), 12.3 (8.1‐57.8), and 11.1 (7.4‐36.5), respectively.

Figure 2.

The boxplots of COL4, MMP‐2, and TNF‐α showing the intensity levels of the 3 proteins at the 5 placement regions. The bottom, middle, and top lines of each box correspond to the 25th percentile, the 50th percentile (median), and the 75th percentile, respectively. The y‐axis represents the intensity level of proteins, which can be captured by the nitrocellulose membrane

3.3. The bilateral difference in the forearms and lower legs

The Bland–Altman plots showed bilateral differences in the intensity levels of 3 proteins in the forearms and lower legs (Figure 3). There was no significant difference between right and left secretion levels on the forearms among COL4 (P = .53), MMP‐2 (P = .68), and TNF‐α (P = .82). There was also no significant difference between right and left secretion levels on the lower legs among COL4 (P = .59), MMP‐2 (P = .18), and TNF‐α (P = .42). In the protein secretion on the forearms, the right side was correlated with the left side in MMP‐2 (r = 0.46, P < .01) and TNF‐α (r = 0.55, P < .01) but not with COL4 (r = 0.14, P = .25). In the protein secretion on the lower legs, the right side was correlated with the left side in COL4 (r = 0.54, P < .01), MMP‐2 (r = 0.67, P < .01), and TNF‐α (r = 0.32, P < .01).

Figure 3.

The upper Bland–Altman plots show the intensity level of the 3 soluble proteins ([A] COL4, [B] MMP‐2, and [C] TNF‐α) on the right and left forearms, respectively. The lower Bland–Altman plots show the intensity level of the 3 soluble proteins ([D] COL4, [E] MMP‐2, and [F] TNF‐α) on the right and left lower legs. The y‐axis and x‐axis indicate the differences and mean between right and left intensity level, respectively

3.4. The relationship between 3 proteins associated with skin tears and systemic factors

In Table 2, lower age (P = .020) was found to be independently associated with higher COL4. Antiplatelet use (P = .009) was found to be independently associated with higher MMP‐2. There were no significant correlations between the systemic factors and TNF‐α.

Table 2.

Multiple regression analysis of systemic factors associated with the intensity of blotted membranes of the 3 soluble proteins

Variables	COL4			MMP‐2			TNF‐α
	β b	P‐valuea	P‐valueb	β b	P‐valuea	P‐value b	β b	P‐valuea	P‐valueb
Age	−.390	.020	.003	—	—	—	—	—	—
Gender	—	—	—	−.073	.170	.592	—	—	—
Body mass index	−.227	.067	.077	—	—	—	—	—	—
Steroid use	—	—	—	—	—	—	.146	.196	.246
Anticoagulants use	.245	.098	.056	—	—	—	—	—	—
Antiplatelet use	—	—	—	.383	.009	.007	−.099	.181	.454
Number of medicines	—	—	—	.201	.049	.184	−.097	.182	.519
Tube feeding	—	—	—	.229	.060	.130	—	—	—
Intravenous feeding	—	—	—	.000	.058	1.000	.077	.193	.594
Oral intake	—	—	—	—	—	—	—	—	—
Diabetes mellitus	—	—	—	—	—	—	—	—	—
Renal failure	—	—	—	—	—	—	—	—	—
Hypertension	—	—	—	—	—	—	—	—	—
Adjusted R 2		.238			.218			.007

COL4, collagen type IV; MMP‐2, matrix metalloproteinase‐2; TNF‐α, tumour necrosis factor‐α.

n = 70.

^aUnivariate analysis.

^bMultiple regression analysis.

4. DISCUSSION

To our knowledge, this clinical study represents the first effort to verify the bilateral comparison of protein secretion on the intact skin of extremities and the relationship between systemic factors related to skin tear development and 3 marker proteins by skin blotting among Japanese elderly patients in a long‐term medical facility.

There was fewer errors of median intensities among examination sites in all of markers. An intensity level of COL4 and MMP‐2 on the extremities, which had high maximum, provided the distribution variations compared with abdomen. This is because skin conditions on the extremities may be because of the influence of systemic factors plus local factors. Several findings associated with systemic factors in COL4 and MMP‐2 were also demonstrated at the abdomen site, which was suitable for measuring systemic factors. In this study, the structure of a protein in addition to examination sites may also have had an impact. COL4 is structural protein contributing to the construction of the extracellular matrix. Therefore, the intensity levels of COL4 at the abdomen, capable of minimising influences by external stimulus, may have been less variable and association with aging. On the other hand, in TNF‐α, the intensity level at the abdomen was the highest among the 5 regions, which had no association with systemic factors. Intensity levels of TNF‐α as water‐soluble proteins may vary widely at the abdomen because of protein secretion from elsewhere in the body. It appears that there are limitations of evaluating skin conditions related to inflammatory marker on the intact skin.

Bland–Altman plots showed no significant difference in the intensities of all 3 soluble proteins between right, and left skin properties on the extremities, following our hypothesis. In addition, right intensity levels on the extremities in 3 proteins had moderate correlation with right intensity levels. This means that the skin blotting method demonstrated an acceptable reliability in assessing the skin properties between the right and left side on the extremities. It appears that collecting samples either from the right or the left skin would be sufficient if skin properties between arms and legs are evaluated using skin blotting. In addition, the distribution of intensity levels between the right and left side of the extremities in all soluble proteins may show an individual difference of skin properties. Thus, collecting samples from both right and left side on the arms or legs and comparing them would contribute to identifying the local risk factors.

We found that the COL4 intensity levels were associated with a decrease in aging (Table 2). The association of the decrease in COL4 within skin in aging is consistent with the findings of the majority of earlier studies investigating skin biopsy,10 and this result followed our hypothesis. We also identified higher intensity levels of MMP‐2 associated with using antiplatelets (Table 2). Antithrombotic therapy is regarded as an essential preventive therapy for recurrent stroke in the guidelines of various associations.25, 26, 27 MMP‐2 and MMP‐9 can cause plaque instability through actions such as matrix degradation, apoptosis of smooth muscle cells, and promotion of inflammatory cell infiltration.28, 29 Their plasma concentrations have been shown to increase after ischaemic stroke.30 In the brain, MMP‐2 level increases after stroke, and its expression is correlated with the extent of neuronal injury, while MMP‐9 level is increased in haemorrhagic transformation after infarction.31 MMP levels may also vary depending on the nature of atherosclerotic plaque. Alternatively, as MMPs play a role in the repair phase of stroke, through modulating brain matrix, brain plasticity, and vascular remodelling,28, 32 their plasma levels may change in relation to the recovery status,30 which may differ according to the location of atherosclerosis. Thus, it follows that the detection of higher levels of MMP‐2 within the skin may be related to the nature of atherosclerotic plaque or recovery status in ischaemic stroke with antiplatelet agents.

Our study has some limitations. The principle of skin blotting is based on the mild, temporal disruption of the skin barrier by local maceration, which allows protein to leak from the epidermis and dermis. The intensity levels of 3 proteins as outcomes of skin blotting may be affected by individual skin barrier function, which consists of sebum, intercellular lipids in the corneal layer of the epidermis, and keratinocyte tight junctions. Skin barrier function, which was decreased by chronological aging and photo aging, can influence the sampling efficiency of proteins using skin blotting. The intensity level of some soluble proteins in elderly skin should be considered with individual difference related to skin barrier function. Therefore, associations between skin barrier function and sampling efficiency of proteins using skin blotting should be identified in the future.

5. CONCLUSION

Collecting samples from either the right or left skin would be sufficient to evaluate skin conditions on the arms or legs in the skin blotting method. Moreover, collecting both right and left samples on the extremities may contribute to identifying the specific local risk factor. Skin blotting was a validated tool to assess skin properties related to systemic factors as well as local factors depending on sampling region. As identifying the local risk factor associated with developing skin disease or injury is important, skin blotting will help in the early detection of local risk factor and identification of the population at high risk of skin disease, all of which can improve the quality of skin management in clinical settings.

Koyano Y, Nakagami G, Minematsu T, Sanada H. Reliability of the skin blotting method when used on the elderly. Int Wound J. 2018;15:807–813. 10.1111/iwj.12931