Sage Journals: Discover world-class research

Abstract

Background:

Skin autofluorescence has been used to assess longer term glycemic control and risk of complications. There is however no agreed site at which autofluorescence should be measured. This study evaluated the within- and between-site agreement in measurement of skin autofluorescence using a noninvasive advanced glycation end product (AGE) reader.

Methods:

Overall, 132 participants were included: 16 with diabetes-related foot ulcers (DFU), 63 with diabetes but without foot ulcers (DMC), 53 without diabetes or foot ulcers (HC). Skin autofluorescence was measured using the AGE Reader (DiagnOptics technologies BV, the Netherlands). Three consecutive skin autofluorescence measurements were each performed at six different body sites: the volar surfaces of both forearms (arms), dorsal surfaces of both calves (legs), and plantar surfaces of both feet (feet). Within- and between-site agreements were analyzed with concordance correlation coefficients (CCC) and 95% confidence intervals (95% CI), absolute mean differences (±standard deviation), and Bland-Altman limits of agreement.

Results:

The agreement between repeat assessments at the same site was almost perfect (CCC [95% CI] ranging from 0.94 [0.91-0.96] for assessments in the right foot to 0.99 [0.99-0.99] for assessments in the left arm). The limits of agreement were narrow within ±0.5 arbitrary units for all sites. The between-site agreement in measurements was poor (CCC < 0.65) with large maximum absolute mean differences (±SD) in arbitrary units (DFU = 3.40 [±2.04]; DMC = 3.15 [±2.45]; HC = 2.72 [±1.83]) and wide limits of agreement.

Conclusions:

Skin autofluorescence measurements can be repeated at the same site with adequate repeatability but measurements at different sites in the same patient have marked differences. The reason for this variation across sites and whether this has any role in diabetes-related complications needs further investigation.

Keywords

advanced glycation end products diabetes mellitus diabetic foot measurement variability tissue glycation skin autofluorescence

Diabetes mellitus can result in both macro-vascular and micro-vascular complications including cardiovascular, eye, kidney and foot diseases.^1-3 Diabetes-related foot disease (DFD) is one of the top ten global causes of disability and is associated with high mortality, frequent recurrence and substantial health care costs.^4,5 Innovative assessment and management strategies are required to combat the global DFD epidemic.⁶

Chronic hyperglycemia due to diabetes is thought to be an underlying cause of diabetic peripheral neuropathy (DPN) and peripheral artery disease (PAD), both of which are significant risk factors for DFD.⁷ Hyperglycemia promotes accumulation of advanced glycation end products (AGEs), which stimulate secondary detrimental effects at a molecular and cellular level.⁸ These include stimulation of the receptor for advanced glycation end products (RAGE) and promotion of inflammation, for example, by up-regulation of nuclear factor-κB and its target genes.⁹ Glycation may also change the biomechanical properties of tissues via collagen cross-linking and thereby influence the mechanical response of tissues to load.^10-12 This heightens the risk of tissue injury due to mechanical stress.^13-16 Hence tissue glycation and repetitive mechanical loading, in the presence of loss of protective sensation due to DPN may promote the development of diabetes-related foot ulcers (DFUs).¹⁷

Autofluorescence of the volar surface of the arm was previously correlated with biopsied tissue concentrations of some AGEs (r = 0.47 and 0.55 for Nε-(carboxyethyl)lysine and pentosidine, respectively).⁸ A large number of studies have demonstrated that higher skin autofluorescence readings are strongly associated with the risk of developing diabetes complications such as DFD,^18-22 cardiovascular events,^22-26 chronic renal disease,^22,27 eye complications,^28,29 and PAD.^30,31 Therefore, skin autofluorescence readings have been suggested to have independent predictive value for diabetes-related complications³² with potential to be used for screening people at risk of developing DFD.

Skin autofluorescence has typically been measured on the volar surface of the forearm,⁸ although measurements performed at the leg have also been reported.^8,33 Despite this, it is currently unclear whether measurements of autofluorescence performed in the feet and legs agree with those performed in the arms. As fluorophores differ between body sites,³⁴ it is likely that skin autofluorescence measurements might also vary.³³ It is possible that the assessment of skin autofluorescence in the feet and legs could provide a more relevant estimate of the effect of glycation on lower limb biomechanics and risk of DFU development. The aim of this study was to investigate the within- and between-site agreement in autofluorescence measurements performed at different skin sites in people with diabetes-related foot ulcers, people with diabetes without foot ulcers and healthy controls.

Methods

Study Design and Setting

Data for this study are from 146 participants recruited for a larger original case-control study performed in North Queensland, Australia.³⁵ We investigated multiple hypotheses in the original study^36,37 and a detailed study protocol exists.³⁵ In brief, there were three groups of eligible participants recruited: people with type 2 diabetes with an active plantar neuropathic foot ulcer (cases; DFU group); people with type 2 diabetes without a history of foot ulcers (diabetes controls; DMC group); and people with no history of type 2 diabetes or a foot ulcer (healthy controls; HC group). The study was approved by two local human research ethics committees. Written informed consent was obtained from all participants.

Inclusion and Exclusion Criteria for Cases and Controls

Inclusion criteria for the DFU group were adults with a confirmed diagnosis of type 2 diabetes and a single active unilateral plantar DFU of longer than 3 months duration.³⁵ The DMC group comprised of adults with a confirmed diagnosis of type 2 diabetes without a history of DFU.³⁵ The HC group comprised of adults without a history of diabetes or DFUs.³⁵ Exclusion criteria included³⁶ (1) orthopedic, musculoskeletal, vestibular, visual or neurological problems affecting mobility (other than DPN); (2) previous orthopedic surgical intervention of the lower limb; (3) PAD defined as an ankle brachial pressure index (ABPI) of < 0.8; and (4) planned vascular reconstructions in the subsequent 12 months.³⁵

Description of Variables Collected

The procedures used for the collection of all variables have been described in detail previously.^35,38 One trained investigator (MEF) conducted all assessments. Skin autofluorescence was measured noninvasively using the AGE Reader (DiagnOptics technologies BV, Groningen, the Netherlands).²⁰ Skin autofluorescence was defined as the average light intensity of the emission spectrum between 420 and 600 nm divided by the average light intensity of the excitation spectrum between 300 and 420 nm, multiplied by 100 and expressed as a percentage in arbitrary units.²⁰

Detailed methodology for collected variables are reported in the main study protocol.³⁵ These variables included age, sex, ethnicity, smoking status, and medical history in all participants and diabetes duration and insulin use in people with diabetes. Physical examination involved the assessment of height, mass, body mass index (BMI), body fat percentage, waist and hip circumference, resting systolic and diastolic blood pressure, 10g monofilament sensation assessed in 10 sites in each foot, and the Michigan Neuropathy score.³⁹ DPN was defined as a monofilament score of ≤8 in either foot. All methods used to obtain clinical and anthopometric measures in the study had good to excellent repeatability.³⁸ All participants with diabetes were referred for a fasting peripheral venous blood sample for the evaluation of glycated hemoglobin A1c (HbA1c), total cholesterol to HDL ratio, C-reactive protein, homocysteine, and estimated glomerular filtration rate (eGFR).⁴⁰

Skin Autofluorescence Measurement Protocol

The AGE Reader illuminates a skin surface of approximately 4 cm², with an excitation light source with peak intensity at 370 nm.^20,21 Guarded against surrounding light, emission light, and reflected excitation light from the skin is measured with a spectrometer. The device was rested on a stable bench top for performing arm measurements and on a raised wooden platform for performing lower limb measurements (see Figure 1). Three consecutive measurements of skin autofluorescence were each performed at six different body sites: arms—the volar aspect of both forearms (approximately 10 cm from the elbow); legs—the exact midpoint between the popliteal fossa and the gastrocnemius muscle belly of both lower legs (to lessen interference from muscle and adipose tissue); and feet—the plantar surface of both forefeet approximately 1 cm distal to the third metatarsal head. Measurements were performed on the same day in a random order. All measurements were performed in a dimly lit testing environment at room temperature to minimize interference from external light sources.⁸

Figure 1.

Skin autofluorescence measurement of the foot. Figure demonstrates the positioning of the foot over the top of the AGE Reader (DiagnOptics technologies BV, Groningen, the Netherlands) to obtain skin autofluorescence measurements. The room was lit for purposes of illustration. Important to note that minimal weight was placed over the top of the device during measurement.

Measurements were performed at sites that were free of visible blood vessels, scars, lichenification, tattoos, or other skin abnormalities including DFUs and callus.⁸ In people with DFUs, the plantar ulcers were located at the heel (n = 2), medial longitudinal arch (n = 3), first metatarsal head (n = 2), fifth metatarsal head (n = 1), apex of digits (n = 3), and hallux (n = 5). None of the DFUs were located at the site of skin autofluorescence measurement at the third metatarsal. Participants were advised to not use lotions, skin products or emollients on their skin at least two days prior to the assessment.⁴¹ Most participants (n = 129) had relatively light skin pigmentation (Fitzpatrick Grades 1 to 3),⁴² two participants had slightly pigmented skin (Fitzpatrick Grade 4), and one participant (n = 1) had deeply pigmented skin (Fitzpatrick Grade 5).⁴² Despite this, measurements were obtained from all three sites in this participant.⁴³

Statistical Analyses

The Statistical Package for the Social Sciences (SPSS) for Windows (released 2013, IBM SPSS Statistics for Windows, Version 22.0, Armonk, NY, IBM Corp) and R (R Core Team, 2015) software (2015; RStudio, Inc., Boston, MA http://www.rstudio.com/) were used for all analyses. Continuous data were reported as mean and standard deviation (SD) if considered to be approximately normally distributed or median and interquartile range (IQR) if skewed. Categorical variables were reported as numbers and percentages (%). Differences in descriptive outcomes between the three groups were compared using a one-way analysis of variance (ANOVA) and Sidak’s post hoc tests or the Kruskal-Wallis test and post hoc Mann-Whitney U test. The Pearson’s chi-square test or Fisher’s exact test were used to assess differences in categorical variables. A P value of less than .05 was considered statistically significant throughout.

Lin’s concordance correlation coefficient (CCC) was calculated to examine the repeatability of measurements.⁴⁴ All CCCs were calculated using an online statistical calculator and reported with two-sided 95% confidence intervals (95% CI).⁴⁵ CCCs were interpreted as “almost perfect” (>0.90), “substantial” (>0.8-0.9), “moderate” (0.65-0.8), and “poor” (<0.65).⁴⁶

Absolute agreement between measures at each body site were assessed using the Bland-Altman limits of agreement (2 x the SD of the difference scores).⁴⁷ Absolute mean differences and SD were reported alongside CCCs for between-site comparisons. The median value of three measurements, the 95% CI of median measurement and range in skin autofluorescence measures were reported for the cohort and for each group. Pearson correlation coefficients for the differences in mean values against the mean, associated P values, and CCCs with 95% CIs were reported with the Bland-Altman plots.

Results

Participant Recruitment and Characteristics

A total of 146 recruited participants were invited to participate, but only 132 (90%) attended the additional study visit required for this study and had skin autofluorescence measurements performed. Multiple attempts were made to contact the participants that failed to attend. Some reasons for failure to participate (n = 14) included lack of available time, lack of interest in being involved, missed appointments, and failure to establish contact. Sixty-one participants (41.7%) could not have all autofluorescence measurements performed because no reading could be recorded, including from the right arm (n = 1), left leg (n = 36), right leg (n = 37), left foot (n = 10), and right foot (n = 29) despite multiple attempts to measure.

There were 16 participants in the DFU group, 63 in the DMC group, and 53 in the HC group (see Table 1). The average age of the participants was 61 years (SD ± 10.4 years), and 76 were male (57.6%) and 68 (51.5%) had a history of smoking. The average BMI was 31.6 kg/m² (SD ± 19.9 kg/m²). In participants with diabetes (n = 79), 30 (38%) had DPN. Differences and similarities between the three groups of participants are detailed in Table 1.

Table 1.

Description of the Cohort Stratified by Group.

Variable	Entire cohort (N = 132)	DFU (n = 16)	DMC (n = 63)	HC (n = 53)	P value
Age (years)	61.0 ± 10.4	63.4 ± 11.3	63.3 ± 9.6^b	57.7 ± 10.2	.008
Males	76 (57.6%)	10 (62.5%)	44 (69.8%)	22 (41.5%)	.008
Ethnicity					.393
Caucasian	129 (97.7%)	16 (100.0%)	62 (98.4%)	51 (96.2%)
Australian Aboriginal/Indigenous/Torres-strait Islander	2 (1.5%)	—	—	2 (3.8%)
Other	1 (0.8%)	—	1 (1.6%)	—
Diabetes duration (years)^#	11.4 ± 8.5*	15.1 ± 7.1	10.8 ± 8.6	—	.054
HbA1c (mmol/mol)^#	55.2 ± 14.2*	57.0 ± 16.8	54.8 ± 13.6	—	.600
Total cholesterol: HDL ratio^#	3.8 ± 1.1*	4.1 ± 1.1	3.8 ± 1.1	—	.375
Homocysteine (mg/L)^#	11.0 ± 3.4*	11.7 ± 4.3	10.8 ± 3.2	—	.365
C-reactive protein (CRP) (mg/L)^#	3.0 ± 3.9*	4.1 ± 4.5	2.8 ± 3.8	—	.250
eGFR mL/min/1.73m^2#	76.1 ± 17.6*	66.1 ± 18.6	78.5 ± 16.6	—	.017
Uses insulin^#	24 (18.2%)*	8 (50.0%)	16 (25.4%)	—	<.001
Smoking status					.224
Never smoked	64 (48.5%)	11 (68.8%)	29 (46.0%)	24 (45.3%)
Ex-smoker	61 (46.2%)	4 (25.0%)	29 (46.0%)	28 (52.8%)
Current smoker	7 (5.3%)	1 (6.3%)	5 (7.9%)	1 (1.9%)
History of hypertension	68 (51.5%)	14 (87.5%)	41 (65.1%)	13 (24.5%)	<.001
History of dyslipidemia	65 (49.2%)	11 (68.8%)	41 (65.1%)	13 (24.5%)	<.001
History of stroke	3 (2.3%)	1 (6.3%)	2 (3.2%)	—	.272
History of ischemic heart disease	22 (16.7%)	5 (31.3%)	15 (23.8%)	2 (3.8%)	.004
History of chronic heart failure	11 (8.3%)	1 (6.3%)	8 (12.7%)	2 (3.8%)	.212
History of chronic pulmonary disease	19 (14.4%)	4 (25.0%)	11 (17.5%)	4 (7.5%)	.138
History of chronic liver disease	7 (5.3%)	2 (12.5%)	5 (7.9%)	—	.070
History of chronic renal impairment	12 (9.1%)	4 (25.0%)	8 (12.7%)	—	.004
Height (cm)	169.9 ± 10.5	174.5 ± 10.4	169.8 ± 10.9	168.6 ± 9.8	.147
Weight (kg)	91.5 ± 60.0	107.3 ± 24.5	102.4 ± 82.6 ^b	73.8 ± 15.0	.019
BMI (kg/m²)	31.6 ± 19.9	35.4 ± 8.9	35.4 ± 27.4^b	26.0 ± 4.6	.026
Body fat percentage (% bf)	28.2 ± 13.1	30.6 ± 12.6	27.4 ± 12.6	28.3 ± 13.8	.677
Waist circumference (cm)	99.6 ± 17.5	116.5 ± 19.3^a,b	106.8 ± 11.1^b	85.8 ± 13.2	<.001
Blood pressure (systolic) mmHg	140.6 ± 20.4	156.1 ± 19.3^a,b	138.5 ± 20.1^b	137.2 ± 19.3	.030
Blood pressure (diastolic) mmHg	79.8 ± 10.7	80.2 ± 7.7	79.8 ± 12.4	79.7 ± 9.6	.992
ABPI^{^}	1.1 ± 0.2	1.0 ± 0.2	1.1 ± 0.1	1.1 ± 0.2	.865
Monofilament score^#	17 ± 5*	8 ± 6	18 ± 4	—	<.001
MNSI symptom score^#	5 ± 2*	7.0 ± 1.3	4.6 ± 2.2	—	<.001
MNSI physical assessment score^#	3 ± 2*	6.6 ± 1.2	2.5 ± 1.8	—	<.001

All data represent mean ± standard deviation, median [interquartile range] or number, and percentages (%).The reported P values indicate main comparison outcomes from one-way ANOVA or Kruskal-Wallis test between the three groups, Pearson’s chi-square test, or Fisher’s exact test between three groups unless indicated by ^# = DFU vs DMC comparison only. ^a = P < .05 vs DMC group; ^b = P < .05 vs HC group on Sidak post hoc tests or pairwise comparisons from Kruskal-Wallis tests. A significance level of P < .05 was used throughout. Diabetes duration indicates fractions of years living with type 2 diabetes mellitus. ^{^}ABPI values represented in the table are the lowest value from both limbs for diabetes and healthy controls and the ABPI from the ulcerated limb for people in the DFU group. Monofilament score is out of a total of 20, measured at 10 sites for each foot. MNSI scores indicate the total scores from the MNSI in relation to the neuropathy symptom score and physical assessment score.

The descriptive characteristics of participants with missing autofluorescence measurements were evaluated (Supplementary Table 1). No obvious differences were noted between this subgroup and the entire cohort.

Within-Site Agreement

Table 2 shows that the agreement between repeat readings performed at the same site was almost perfect, with CCC ranging from 0.94 [0.91-0.96] for the right foot to 0.99 [0.99-0.99] for the left arm. Supplementary Table 2 shows all three participant groups also had substantial to almost perfect within-site agreement at all six sites, except for the left leg in the DFU group CCC = 0.75 [0.31-0.93]. The Bland-Altman limits of agreement were within ± 0.5 arbitrary units for all sites (Figure 2).

Table 2.

Agreement in Measurements Performed at Each Body Site Stratified by Limb.

	Participants (N=)	CCC measurement 1 vs 2 [95% CI]	CCC measurement 1 vs 3 [95% CI]	CCC measurement 2 vs 3 [95% CI]	Median age reader measurement [IQR]*	Upper and lower 95% CI of median	Range
Left arm	132	0.99 [0.98 to 0.99]	0.98 [0.97-0.99]	0.99 [0.99-0.99]	2.7 [0.8]	[2.6-2.7]	0.3-4.9
Right arm	131	0.97 [0.95 to 0.98]	0.95 [0.93-0.96]	0.99 [0.99-0.99]	2.7 [0.7]	[2.6-2.8]	0.3-4.6
Left leg	94	0.97 [0.96-0.98]	0.97 [0.95-0.98]	0.99 [0.97-0.99]	2.3 [0.8]	[2.2-2.4]	0.3-5.9
Right leg	94	0.99 [0.98-0.99]	0.99 [0.99-0.99]	0.99 [0.99-0.99]	2.4 [0.8]	[2.3-2.5]	0.3-4.6
Left foot	121	0.97 [0.96-0.99]	0.96 [0.94-0.98]	0.98 [0.97-0.98]	4.8 [1.5]	[4.7-5.1]	2.1-9.5
Right foot	103	0.94 [0.91-0.96]	0.97 [0.96-0.98]	0.96 [0.94-0.97]	5.1 [2.4]	[5.0-5.3]	2.0-11.0

Numerical figures indicate CCC measurements and two-sided 95% confidence intervals. CCC values were calculated using http://www.niwa.co.nz/node/104318/concordance. Two-sided 95% CIs are presented with the CCC value for each measurement. The strength of agreement was considered as almost perfect >0.90; substantial >0.8-0.9; moderate 0.65-0.8; and poor <0.65. AGE reader measurement reports the median measurement from three consecutive measurements and the IQR. 95% CIs indicate the 95% CIs for the median. Range indicates the minimum-maximum values for each measurement. CCCs are reported to two decimal points, AGE reader measurements to one decimal point.

Figure 2.

Bland-Altman plots for repeated measures at all six sites. All measurements represent skin autofluorescence measurements in AU. Figure outlines the Bland-Altman plots as represented by means on the x-axis and mean differences on the y-axis between measurement 1 and measurement 3 at each site. The black circles represent measurements in the DFU group (cases), the red circles indicate measurements in the DMC group (control), and the green circles represent measurements in the HC group. The middle dotted line represents the mean difference; the upper and lower dotted lines indicate ± 2 standard deviations from the mean difference (Bland-Altman limits of agreement). The regression line represents a linear regression of the differences and means. Pearson = pearson correlation coefficient; the reported P values represent the associated P values from the correlation analyses. CCC = concordance correlation coefficients and its associated 95% CI.

Agreement Between Left- and Right-Sided Measurements

The agreement based on CCC [95% CI] for each group is reported in Supplementary Table 3 and was moderate to substantial for the arms (0.80 [0.56-0.92] DFU to 0.86 [0.78-0.92] HC), poor to substantial for the legs (0.64 [0.25-0.85]) DFU to 0.82 [0.69-0.90]) HC), and poor to moderate for the feet (0.43 [0.23-0.59] HC to 0.71 [0.33-0.89] DFU). Post hoc analyses suggested that the presence of a DFU and the presence of DPN had minimal impact on right and left side between-site agreement (Supplementary Table 4). The Bland-Altman limits of agreement between measurements performed at right and left limbs are shown in Figure 3 stratified by group. The limits of agreement were larger for the feet (±2.5 arbitrary units) compared to the arms (±0.5 arbitrary units). The absolute mean differences (±SD) ranged from 0.03 (±0.56) to 0.13 (±0.33) arbitrary units for the DFU group and HC group for arm measurements and 0.13 (±2.54) to 0.79 (±2.83) arbitrary units for the HC group and DMC group for feet measurements.

Figure 3.

Bland-Altman plots for left and right measurement agreement. All measurements represent skin autofluorescence measurements in AU. Figure outlines the Bland-Altman plots as represented by means on the x-axis and mean differences on the y-axis between left- and right-sided measurements using averaged values from three measurements. The black circles represent measurements in the DFU group (cases), the red circles indicate measurements in the DMC group (control), and the green circles represent measurements in the HC group. The middle dotted line represents the mean difference; the upper and lower dotted lines indicate ± 2 standard deviations from the mean difference (Bland-Altman limits of agreement). The regression line represents a linear regression of the differences and means. Pearson = pearson correlation coefficient; the reported P values represent the associated P values from the correlation analyses. CCC = concordance correlation coefficients and its associated 95% CI.

Agreement Between Different Body Sites

Mostly there was poor between site agreement based on CCC [95% CI] and absolute mean differences (±SD) (Figure 4, Table 3). The Bland-Altman limits of agreement are shown in Figure 4 and indicated wider ranges for measurements performed in the feet compared to arms and measurements performed in the feet compared to the legs. There was moderate agreement between measurements performed in the left arm and right leg, and the right arm and right leg in the HC group (Table 3). The absolute mean differences were largest for measurements performed in the left leg and left foot in all three groups (3.40 [±2.04] arbitrary units DFU group; 3.15 [±2.45] arbitrary units DMC group; and 2.72 [±1.83] arbitrary units HC group).

Figure 4.

Bland-Altman plots for between site agreement. All measurements represent skin autofluorescence measurements in AU. Figure outlines the Bland-Altman plots as represented by means on the x-axis and mean differences on the y-axis between body sites. The black circles represent measurements in the DFU group (cases), the red circles indicate measurements in the DMC group (control), and the green circles represent measurements in the HC group. The middle-dotted line represents the mean difference; the upper and lower dotted lines indicate ± 2 standard deviations from the mean difference (Bland-Altman limits of agreement). The regression line represents a linear regression of the differences and means. Pearson = pearson correlation coefficient; the reported P values represent the associated P values from the correlation analyses. CCC = concordance correlation coefficients and its associated 95% CI.

Table 3.

Between Site Agreements in Measurements Stratified by Group.

	DFU group		DMC group		HC group
	CCC [95% CI]	Absolute mean difference (SD)	CCC [95% CI]	Absolute mean difference (SD)	CCC [95% CI]	Absolute mean difference (SD)
Left arm vs left leg	0.24[−0.09 to 0.52]	1.54 (1.22)	0.46 [0.21 to 0.66]	1.24 (1.60)	0.47 [0.27 to 0.63]	0.78 (0.94)
Left arm vs left foot	0.10 [−0.12 to 0.31]	1.43 (2.28)	0.01 [0.00 to 0.15]	1.76 (1.93)	−0.01 [−0.01 to −0.05]	1.89 (1.66)
Left arm vs right leg	0.44 [0.07 to 0.70]	0.98 (1.00)	0.22 [−0.09 to 0.49]	1.26 (1.75)	0.68 [0.51 to 0.79]	0.70 (0.95)
Left arm vs right foot	−0.00 [−0.22 to 0.21]	0.97 (2.50)	0.01 [−0.06 to 0.09]	1.04 (2.87)	−0.05 [−0.11 to −0.02]	1.76 (2.94)
Right arm vs right leg	0.51 [0.13 to 0.76]	0.94 (1.21)	0.33 [0.02 to 0.57]	1.20 (1.84)	0.72 [0.55 to 0.83]	0.57 (0.96)
Right arm vs right foot	−0.07 [−0.24 to −0.10]	1.00 (2.76)	0.03 [−0.01 to 0.010]	1.09 (2.89)	−0.02 [−0.07 to 0.30]	1.89 (2.89)
Right arm vs left leg	0.34 [−0.08 to 0.65]	1.51 (1.36)	0.44 [0.17 to 0.65]	1.19 (1.68)	0.53 [0.31 to 0.70]	0.65 (0.97)
Right arm vs left foot	−0.00 [−0.19 to 0.18]	1.47 (2.65)	0.08 [0.01 to 0.15]	1.82 (1.90)	0.01 [−0.04 to 0.06]	2.01 (1.60)
Left leg vs left foot	0.04 [−0.10 to 0.17]	3.40 (2.04)	0.03 [−0.04 to 0.10]	3.15 (2.45)	−0.00 [−0.04 to 0.04]	2.72 (1.83)
Left leg vs right foot	0.06 [−0.12 to 0.23]	3.09 (2.06)	−0.01 [−0.09 to 0.08]	2.51 (2.96)	−0.02 [−0.07 to 0.03]	2.69 (2.85)
Right leg vs right foot	0.13 [−0.16 to 0.39]	2.22 (2.72)	−0.01 [−0.10 to 0.08]	2.58 (3.18)	−0.01 [−0.06 to 0.03]	2.66 (2.76)
Right leg vs left foot	0.09 [−0.16 to 0.33]	2.75 (2.15)	−0.01 [−0.11 to 0.09]	3.22 (2.29)	0.00 [−0.04 to 0.04]	2.64 (1.84)

Discussion

This is the first study to report the agreement between repeat skin autofluorescence measurements at the same and different body sites in a heterogeneous cohort. The agreement in measurements repeated at the same site were substantial to almost perfect in each group. The agreement between left- and right-sided measurements ranged from substantial to poor in the three groups, and the limits of agreement in measurements and absolute mean differences were larger for the feet compared to the arms. Overall there was poor agreement between measurements performed at the six different sites and the lack of between-site agreement seemed to increase with increased distance between sites. These findings suggest that there may be variation in glycation at different sites in the body, although other factors, such as skin translucency, may also explain the results. It is possible but unproven that the assessment of skin autofluorescence in the feet could provide a more accurate prediction of the risk of DFD.

Most previous studies have measured autofluorescence in the upper arm.^32,41,48,49 The differences in measurements performed at different body sites were noted to be larger than differences previously reported between different patient populations with and without diabetes or DFD.^19,22,32,50 This highlights the need for a consistent measurement site. Previous studies have also reported excellent levels of repeatability in skin autofluorescence readings consistently performed at the same site in the arm.^8,51 Low coefficient of variations (6.9% and 10.9%) have previously been reported for skin autofluorescence measurements repeated 6 and 12 weeks apart in the arms of people with type 2 diabetes.⁵² Variations in measurement technique, lack of consistency in the site of measurement, and biological variation over time were suggested to contribute to the measurement variability.⁵² The current study expands on prior research by evaluating Bland-Altman plots and absolute agreement⁵³ and including the legs and feet as additional sites of measurement in both people at risk of developing DFD and people with active DFUs.

Skin autofluorescence readings in the feet were consistently higher than measurements in the legs and arms in all three groups. This finding is consistent with higher levels of glycation in the feet compared to more proximal sites, which could in part explain why DPN mostly affects the lower not upper limbs.⁵⁴ Since our findings indicate autofluorescence values differ substantially between different sides of the body, averaging values from left and right sides of the body is not recommended. The results of this study also suggest that the skin autofluorescence measured in the arm (the commonest site of measurement)^{19,26,28,55-58} are not necessarily representative of measurements at other body sites.

Different body sites have been reported to have different levels of fluorophores which may in part explain the findings of this study. Na and coworkers suggested a poor agreement between skin autofluorescence measurements performed at different sites could be due to differences in the levels of fluorophores within the skin.³⁴ Fluorophores such as nicotinamide adenine dinucleotide (NADH), flavin adenine dinucleotide (FAD), and porphyrins have an excitation spectrum similar to AGEs and differences in their concentration at different body sites may affect skin autofluorescence.⁵⁰ Although up to 76% of the variance in the skin autofluorescence signal has previously been correlated with different AGE concentrations, the AGE reader cannot distinguish between AGEs and other fluorophores in the same excitation spectrum.⁴¹

Yamanaka and colleagues suggested that the presence of veins increases the skin autofluorescence measurements by 1.5-fold and therefore it is possible that variation in the distribution of veins at different body sites may explain some of the measurement variation.⁴⁹ Readings also likely depend on the depth of penetrance of light and the density of skin.⁴⁹ The skin on the plantar surface of the feet has a different tissue composition and thickness to the skin in the arms or legs. This likely affects the penetrance of the laser used to measure skin autofluorescence and may have contributed to the variation in measurements.

The use of multiple analysis methods including Bland-Altman plots with limits of agreement and absolute mean differences are a strength of this study.^47,53 Another strength was the inclusion of a heterogeneous group of people considered to have a wide spectrum of skin autofluorescence measurements and the stratification of the cohort based on group. The age of participants in the DMC group and HC group and the duration of diabetes in the DFU and DMC groups differed and both are known to impact skin autofluorescence readings.²¹ Therefore we evaluated measurement variation in the different participant groups.

The main limitation of this study was the convenience and heterogeneous sample. Skin autofluorescence readings for the lower limb could not be obtained from up to 37% of our cohort and the reasons for this were unclear. Patients with missing data did not have unique characteristics as compared to the overall cohort and the AGE reader data obtained in our study was comparable to other studies in the field.^19,32,50 The study was cross-sectional, and the investigation did not evaluate between-day measurement variability or the relationship between skin autofluorescence and circulating AGEs or HbA1c.^56,59 Last, the focus of this study was not to compare skin autofluorescence between patients that did or did not have DFD. The study was not adequately powered for this comparison that has been previously reported in prior studies.^19,57 Furthermore, this study measured skin autofluorescence, not AGE concentrations, and therefore it is uncertain how the variation of autofluorescence reported here relates to AGE distributions throughout the body.

Conclusion

Skin autofluorescence measurements repeated at the same body site had good agreement whereas agreement between measurements performed at different body sites was poor. Further studies are required to determine the reasons for this variation and the ideal site to perform measurements in people with DFD.

Supplemental Material

Supplementary_Files – Supplemental material for Within- and Between-Body-Site Agreement of Skin Autofluorescence Measurements in People With and Without Diabetes-Related Foot Disease

Supplemental material, Supplementary_Files for Within- and Between-Body-Site Agreement of Skin Autofluorescence Measurements in People With and Without Diabetes-Related Foot Disease by Malindu E. Fernando, Robert G. Crowther, Peter A. Lazzarini, Kunwarjit S. Sangla, Scott Wearing, Petra Buttner and Jonathan Golledge in Journal of Diabetes Science and Technology

Footnotes

Acknowledgements

The authors wish to thank all the participants of the study and the administrative and academic staff at the Movement Analysis Laboratory, Sports and Exercise Science, James Cook University, Townsville, Australia, for assistance with study visits and the utilization of space to conduct the study.

Abbreviations

ABPI, ankle brachial pressure index; AGE, advanced glycation end product; BMI, body mass index; CCC, concordance correlation coefficient; CI, confidence interval; DFD, diabetes-related foot disease; DFU, diabetes foot ulcer; DMC, diabetes mellitus control; DPN, diabetic peripheral neuropathy; eGFR, estimated glomerular filtration rate; HC, healthy control; IQR, interquartile range; MNSI, Michigan Neuropathy Screening Instrument; PAD, peripheral artery disease; RAGE, receptor for advanced glycation end products; SD, standard deviation.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding from the Department of Health, Queensland Government under the Health Practitioner Research Grant Scheme (2013-2014), funding from the Graduate Research School and Strategic Research Investment Fund for The Ulcer and Wound Healing Consortium (UHEAL) at James Cook University, and funding from the National Health and Medical Research Council supported this work. JG holds a Practitioner Fellowship from the National Health and Medical Research Council, Australia (1117061). JG holds a Senior Clinical Research Fellowship from the Queensland Government. MEF was supported by an Australian Postgraduate Award Scholarship at James Cook University and a scholarship from College of Medicine, James Cook University at the time of the study.

ORCID iDs

Malindu E. Fernando

Peter A. Lazzarini

Supplemental Material

Supplemental material for this article is available online.

References

Stratton

Adler

Neil

HAW

et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ. 2000;321:405-412.

The ADVANCE Collaborative Group. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med. 2008;358:2560-2572.

The ACCORD Study Group. Long-term effects of intensive glucose lowering on cardiovascular outcomes. N Engl J Med. 2011;364:818-828.

Lazzarini

Pacella

Armstrong

van Netten

JJ.

Diabetes-related lower-extremity complications are a leading cause of the global burden of disability. Diabet Med. 2018; 35:1297.

Cheng

Lazzarini

Gibb

et al. A cost-effectiveness analysis of optimal care for diabetic foot ulcers in Australia [published online ahead of print August 4, 2016]. Int Wound J. doi:10.1111/iwj.12653.

Jeffcoate

Vileikyte

Boyko

Armstrong

Boulton

AJM

. Current challenges and opportunities in the prevention and management of diabetic foot ulcers. Diabetes Care. 2018;41:645-652.

Fernando

Seneviratne

Tan

et al. Intensive versus conventional glycaemic control for treating diabetic foot ulcers. Cochrane Database Syst Rev. 2016(1):CD010764.

Meerwaldt

Graaff

Oomen

et al. Simple non-invasive assessment of advanced glycation endproduct accumulation. Diabetologia. 2004;47:1324-1330.

Goldin

Beckman

Schmidt

Creager

MA.

Advanced glycation end products: sparking the development of diabetic vascular injury. Circulation. 2006;114:597-605.

10.

Wrobel

Najafi

Diabetic foot biomechanics and gait dysfunction. J Diabetes Sci Technol 2010;1:833-845.

11.

Giacomozzi

D’Ambrogi

Uccioli

Macellari

Does the thickening of Achilles tendon and plantar fascia contribute to the alteration of diabetic foot loading?

Clin Biomech (Bristol, Avon). 2005;20:532-539.

12.

D’Ambrogi

Giurato

D’Agostino

et al. Contribution of plantar fascia to the increased forefoot pressures in diabetic patients. Diabetes Care. 2003;26:1525-1529.

13.

Dinh

Veves

A review of the mechanisms implicated in the pathogenesis of the diabetic foot. Int J Low Extrem Wounds. 2005;4:154-159.

14.

Veves

Murray

Young

Boulton

AJM

. The risk of foot ulceration in diabetic patients with high foot pressure: a prospective study. Diabetologia. 1992;35:660-663.

15.

Caselli

Pham

Giurini

Armstrong

Veves

The forefoot-to-rearfoot plantar pressure ratio is increased in severe diabetic neuropathy and can predict foot ulceration. Diabetes Care. 2002;25:1066-1071.

16.

Fernando

Crowther

Lazzarini

et al. Biomechanical characteristics of peripheral diabetic neuropathy: a systematic review and meta-analysis of findings from the gait cycle, muscle activity and dynamic barefoot plantar pressure. Clin Biomech (Bristol, Avon). 2013;28:831-845.

17.

Huijberts

Schaper

Schalkwijk

CG.

Advanced glycation end products and diabetic foot disease. Diabetes Metab Res Rev. 2008;24(suppl 1):S19-S24.

18.

Vouillarmet

Maucort-Boulch

Michon

Thivolet

Advanced glycation end products assessed by skin autofluorescence: a new marker of diabetic foot ulceration. Diabetes Technol Ther. 2013;15:601-605.

19.

Han

Huang

Smit

AJ.

Elevated skin autofluorescence is strongly associated with foot ulcers in patients with diabetes: a cross-sectional, observational study of Chinese subjects. J Zhejiang Univ Sci B. 2012;13:372-377.

20.

Gerrits

Lutgers

Kleefstra

et al. Skin autofluorescence: a tool to identify type 2 diabetic patients at risk for developing microvascular complications. Diabetes Care. 2008;31:517-521.

21.

Meerwaldt

Links

Graaff

et al. Increased accumulation of skin advanced glycation end-products precedes and correlates with clinical manifestation of diabetic neuropathy. Diabetologia. 2005;48:1637-1644.

22.

Noordzij

Mulder

Oomen

et al. Skin autofluorescence and risk of micro- and macrovascular complications in patients with type 2 diabetes mellitus—a multi-centre study. Diabet Med. 2012;29:1556-1561.

23.

Lutgers

Gerrits

Graaff

et al. Skin autofluorescence provides additional information to the UK Prospective Diabetes Study (UKPDS) risk score for the estimation of cardiovascular prognosis in type 2 diabetes mellitus. Diabetologia. 2009;52:789-797.

24.

Onofriescu

Voroneanu

Hogas

et al. Evaluation of arterial stiffness and tissue advanced glycation end products by skin autofluorescence in peritoneal dialysis patients. Paper presented at: NDT Plus Conference; 2010; Munich, Germany.

25.

Schmidt

AM.

Skin autofluorescence, 5-year mortality, and cardiovascular events in peripheral arterial disease: all that glitters is surely not gold. Arterioscler Thromb Vasc Biol. 2014;34:697-699.

26.

Tanaka

Tani

Asai

et al. Skin autofluorescence is associated with severity of vascular complications in Japanese patients with type 2 diabetes. Diabet Med. 2012;29:492-500.

27.

Smit

Gerrits

EG.

Skin autofluorescence as a measure of advanced glycation endproduct deposition: a novel risk marker in chronic kidney disease. Curr Opin Nephrol Hypertens. 2010;19:527-533.

28.

Yasuda

Shimura

Kunikata

et al. Relationship of skin autofluorescence to severity of retinopathy in type 2 diabetes. Curr Eye Res. 2015;40:338-345.

29.

Hirano

Iesato

Toriyama

Imai

Chiba

Murata

Correlation between diabetic retinopathy severity and elevated skin autofluorescence as a marker of advanced glycation end-product accumulation in type 2 diabetic patients. J Diabetes Complications. 2014;28:729-734.

30.

de Vos

Lefrandt

Dullaart

RPF

Zeebregts

Smit

AJ.

Advanced glycation end products: an emerging biomarker for adverse outcome in patients with peripheral artery disease. Atherosclerosis. 2016;254:291-299.

31.

de Vos

Noordzij

Mulder

et al. Skin autofluorescence as a measure of advanced glycation end products deposition is elevated in peripheral artery disease. Arterioscler Thromb Vasc Biol. 2013;33:131-138.

32.

Osawa

Katakami

Sato

et al. Skin autofluorescence is associated with vascular complications in patients with type 2 diabetes. J Diabetes Complications. 2018;32:839-844.

33.

Blaauw

Smit

van Pampus

et al. Skin autofluorescence, a marker of advanced glycation end products and oxidative stress, is increased in recently preeclamptic women. Am J Obstet Gynecol. 2006;195:717-722.

34.

Stender

Wulf

HC.

Autofluorescence spectrum of skin: component bands and body site variations. Skin Res Technol. 2000;6:112-117.

35.

Fernando

Crowther

Cunningham

Lazzarini

Sangla

Golledge

Lower limb biomechanical characteristics of patients with neuropathic diabetic foot ulcers: the diabetes foot ulcer study protocol. BMC Endocr Disord. 2015;15:59.

36.

Fernando

Crowther

Lazzarini

Sangla

Buttner

Golledge

Gait parameters of people with diabetes-related neuropathic plantar foot ulcers. Clin Biomech (Bristol, Avon). 2016;37:98-107.

37.

Fernando

Crowther

Lazzarini

et al. Plantar pressures are higher in cases with diabetic foot ulcers compared to controls despite a longer stance phase duration. BMC Endocr Disord. 2016;16:1-10.

38.

Fernando

Crowther

Cunningham

et al. The reproducibility of acquiring three dimensional gait and plantar pressure data using established protocols in participants with and without type 2 diabetes and foot ulcers. J Foot Ankle Res. 2016;9:4.

39.

Michigan Diabetes Research and Training Center. Michigan Neuropathy Screening Instrument. University of Michigan Health System; 2013. Available at: www.med.umich.edu/borc/profs/documents/svi/MNSI_patient.pdf.

40.

Golledge

Ewels

Muller

Walker

PJ.

Association of chronic kidney disease categories defined with different formulae with major adverse events in patients with peripheral vascular disease. Atherosclerosis. 2014;232:289-297.

41.

Da Moura Semedo

Webb

Waller

Khunti

Davies

Skin autofluorescence, a non-invasive marker of advanced glycation end products: clinical relevance and limitations. Postgrad Med J. 2017;93:289-294.

42.

Fitzpatrick

TB.

The validity and practicality of sun-reactive skin types I through VI. Arch Dermatol. 1988;124:869-871.

43.

den Engelsen

van den Donk

Gorter

Salomé

Rutten

GE.

Advanced glycation end products measured by skin autofluorescence in a population with central obesity. Dermatoendocrinol. 2012;4:33-38.

44.

Lin

LI-K

. A concordance correlation coefficient to evaluate reproducibility. Biometrics. 1989;45:255-268.

45.

NIWA. Statistical calculators; 2013. Available at: http://www.niwa.co.nz/node/104318/concordance.

46.

McBride

GB.

A proposal for strength-of-agreement criteria for Lin’s Concordance Correlation Coefficient Hamilton, New Zealand;

2005. Available at: https://www.medcalc.org/download/pdf/McBride2005.pdf.

47.

Altman

Bland

JM.

Measurement in medicine: the analysis of method comparison studies. Statistician. 1983;32:307-317.

48.

Cavero-Redondo

Soriano-Cano

Alvarez-Bueno

et al. Skin autofluorescence-indicated advanced glycation end products as predictors of cardiovascular and all-cause mortality in high-risk subjects: a systematic review and meta-analysis. J Am Heart Assoc. 2018;7:e009833.

49.

Yamanaka

Matsumura

Ohno

R-I

et al. Non-invasive measurement of skin autofluorescence to evaluate diabetic complications. J Clin Biochem Nutr. 2016;58:135-140.

50.

Lutgers

Graaff

Links

et al. Skin autofluorescence as a noninvasive marker of vascular damage in patients with type 2 diabetes. Diabetes Care. 2006;29:2654-2659.

51.

Meerwaldt

Hartog

JWL

Graaff

et al. Skin Autofluorescence, a measure of cumulative metabolic stress and advanced glycation end products, predicts mortality in hemodialysis patients. J Am Soc Nephrol. 2005;16:3687-3693.

52.

Stirban

Pop

Fischer

Heckermann

Tschoepe

Variability of skin autofluorescence measurement over 6 and 12 weeks and the influence of benfotiamine treatment. Diabetes Technol Ther. 2013;15:733-737.

53.

Bland

Altman

DG.

Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1:307-310.

54.

Tesfaye

Boulton

Dyck

et al. Diabetic neuropathies: update on definitions, diagnostic criteria, estimation of severity, and treatments. Diabetes Care. 2010;33:2285-2293.

55.

van Waateringe

Fokkens

Slagter

et al. Skin autofluorescence predicts incident type 2 diabetes, cardiovascular disease and mortality in the general population. Diabetologia. 2019;62:269-280.

56.

Hangai

Takebe

Honma

et al. Association of advanced glycation end products with coronary artery calcification in Japanese subjects with type 2 diabetes as assessed by skin autofluorescence. J Atheroscler Thromb. 2016;23:1178-1187.

57.

Liu

Gao

et al. The association between skin autofluorescence and vascular complications in Chinese patients with diabetic foot ulcer: an observational study done in Shanghai. Int J Low Extrem Wounds. 2015;14:28-36.

58.

McIntyre

Fluck

McIntyre

Taal

MW.

Skin autofluorescence and the association with renal and cardiovascular risk factors in chronic kidney disease stage 3. Clin J Am Soc Nephrol. 2011;6:2356-2363.

59.

Gerrits

Lutgers

Kleefstra

et al. Skin advanced glycation end product accumulation is poorly reflected by glycemic control in type 2 diabetic patients (ZODIAC-9). J Diabetes Sci Technol. 2008;2:572-577.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.19 MB