Abstract
Purpose
To compare tissue sodium and fat content in the upper and lower extremities of participants with lipedema vs. controls using magnetic resonance imaging (MRI).
Methods
MRI was performed at 3.0T in females with lipedema (n=15, age=43.2±10.0 years, BMI=30.3±4.4 kg/m2) and controls without lipedema (n=14, age=42.8±13.2 years, BMI=28.8±4.4 kg/m2). Participants were assessed for pain and disease stage. Sodium MRI was performed in the forearm and calf to quantify regional tissue sodium content (TSC, mmol/L). Chemical-shift-encoded water-fat MRI was performed in identical regions for measurement of fat/water (ratio).
Results
In the calf, skin TSC (16.3±2.6 vs. 14.4±2.2 mmol/L, p=0.04), muscle TSC (20.3±3.0 vs. 18.3±1.7 mmol/L, p=0.03), and fat/water (1.03±0.37 vs. 0.56±0.21 ratio, p<0.001) were significantly higher in lipedema vs. control participants. In the forearm, skin TSC (13.4±3.3 vs. 12.0±2.3 mmol/L, p=0.2, Cohen’s d=0.50) and fat/water (0.65±0.24 vs. 0.48±0.24 ratio, p=0.07, Cohen’s d=0.68) demonstrated moderate effect sizes in lipedema vs. control participants. Calf skin TSC was significantly correlated with pain (Spearman’s rho=0.55, p=0.03) and disease stage (Spearman’s rho=0.82, p<0.001) among participants with lipedema.
Conclusion
MRI of tissue sodium and fat content are significantly higher in the lower extremities, but not upper extremities, of patients with lipedema compared to BMI-matched controls.
Keywords: lipedema, lipoedema, MRI, body composition, sodium, CSE MRI, adipose tissue disorder
Introduction
Lipedema is an adipose tissue disorder that almost exclusively affects women and is characterized by excessive subcutaneous adipose tissue accumulation. Lipedema is often misdiagnosed as obesity1,2, a clinical challenge that is receiving increased attention3,4. Adipose tissue deposition in affected areas is resistant to common diet and exercise treatments2, and even bariatric surgery5. Many women remain undiagnosed or only receive a diagnosis once adipose accumulation becomes advanced6. As the condition progresses, lipedema is often associated with local pain, bruising, and swelling.
Lipedema diagnosis, and corresponding treatment development, is limited by a lack of objective metrics to quantify tissue profiles of lipedema which would be required to distinguish lipedema from other common conditions such as obesity. To address this limitation, noninvasive magnetic resonance imaging (MRI) techniques are being investigated to examine tissue composition in patients with lipedema. Prior studies have demonstrated lower extremity tissue edema in lipedema case examples using T1/T2-weighted MRI7,8. Recent work has focused on characterizing sodium and fat content using multi-nuclear sodium and proton (23Na/1H) MRI9. In this work, it was shown that lipedema patients have elevated tissue sodium and fat content in the calf compared to controls matched for age, sex, body-mass-index (BMI), race, and calf circumference. These findings indicate that MRI measurement of tissue sodium and fat content may help distinguish patients with lipedema from those with obesity, and provide a foundation for developing objective diagnostic criteria.
Prior work focused only on the lower extremities, yet lipedema is reported to affect the upper extremities in approximately 80% of patients6, albeit not as severely as the lower extremities. It remains unclear whether elevated tissue sodium and fat are present in other affected regions, such as the upper extremities, or whether these imaging signatures are primarily elevated in the legs. Characterizing upper and lower extremity disease pathophysiology could indicate whether systemic or local treatments will better manage lipedema symptoms, and would provide insight into disease mechanisms.
Accordingly, the goal of this work was to extend prior sodium MRI and chemical-shift-encoded (CSE) water-fat MRI methods to lower and upper extremities in patients with lipedema. We tested the hypothesis that MRI measures of tissue sodium and fat content are higher in the lower extremities, but not upper extremities, of patients with arm and leg lipedema compared to controls. Secondary analyses evaluated the distribution of tissue sodium and fat content in the upper vs. lower extremities within each group, and the relationship between imaging metrics and clinical indicators of disease severity such as leg pain and disease stage.
Methods
Volunteer inclusion and exclusion criteria
All volunteers (n=29) provided informed consent in accordance with the Vanderbilt University Medical Center Institutional Review Board. Volunteers with lipedema (n=15) were recruited through the Vanderbilt INFORM database, flyer distribution, ReseachMatch.org, and through an enrollment survey distributed through social media support groups. Control volunteers (n=14) were recruited using the same recruitment mechanisms to match the sex, race, age, and BMI of lipedema participants. The survey was disseminated through REDCap™ and included questions regarding lipedema symptoms, medical history related to inclusion/exclusion criteria, and the visual analogue scale (VAS, presented electronically as a sliding scale from 0–100) regarding the experience of pain in the legs on a normal day.
Participants with lipedema were required to meet all primary lipedema inclusion criteria1: bilateral swelling of the legs, negative Stemner’s sign, and positive for stage and type of lipedema; at least one secondary criterion was required: lower extremity pain, family history of lipedema, non-pitting lower extremity edema, easy bruising, or hypermobility. To test our study hypothesis regarding sodium and fat content in the upper and lower extremities affected by lipedema, participants with lipedema were also required to have arm involvement, which is representative of approximately 80% of patients with lipedema6. Exclusion criteria for both patients and controls were prior liposuction to the calves or arms, current skin infections or other signs of acute inflammation, and history of uncontrolled hypertension, diabetes, or arthritis.
Physical exam
On the day of the study visit, a physical exam was performed to measure height, weight, and blood pressure (BP). Each limb was also assessed using a perometer (400NT, Pero-System Messgeraete GmbH, Wuppertal, Germany) to measure upper and lower extremity volume (mL) and circumference (cm). Lipedema participants were assessed by a certified LANA (Lymphology Association of North America) lymphedema therapist (P.M.C.D.) to confirm they met inclusion criteria. Participants with lipedema were assessed for disease stage, either: the legs have normal skin appearance over a thickened and nodular hypodermis with a soft tissue turgor possibly tender to touch (stage 1), the involved areas have dimpling of the skin and potentially larger palpable nodules in the thickened adipose tissue resulting in an irregular tissue texture (stage 2), or the involved areas are predominantly distorted by enlarged lobules of tissue that may compromise joint integrity and impair mobility and balance (stage 3)1,6. Participants with lipedema were assessed for lipedema type, either involvement of the hips and buttocks (type 1), hips to knees (type 2), or hips to ankles (type 3). Additionally, arm involvement was assessed as present if the upper extremity exhibited subcutaneous tissue thickening, soft tissue turgor, and nodular palpation of the hypodermis10.
Magnetic resonance imaging (MRI)
Participants underwent an MRI exam on a 3.0T scanner (Philips Healthcare, Best, The Netherlands). The exam consisted of multi-nuclear sodium imaging (23Na-MRI) using a single-tuned receive-only quadrature sodium coil (Rapid Biomedical GmbH, Rimpar, Germany), and proton (1H-MRI) imaging of the upper and lower extremities (Figure 1). Participants were positioned supine to image the calf at the mid gastrocnemius muscle, and prone with the arm extended overhead to image the forearm at approximately 5 cm distal to the antecubital fossa. Imaging was performed on either the right or left side for both the calf and forearm, in order to (i) exclude scanning any recent injuries to either limb, (ii) allow for the arm to be raised above the head, and (iii) give preference to the limb that experiences greater pain. The right side was chosen if all of these considerations were negative for a limb preference. Overall, the right limbs were scanned in 19 participants, and left limbs in 10 participants.
Figure 1.
Noninvasive MRI was performed in the upper and lower extremities. Example images of a participant with lipedema display a) anatomical whole-body CSE fat-weighted imaging, sodium MRI (colored images, middle) and CSE water-weighted MRI (right) in identical slices of the b-c) forearm and d-e) calf. The water-weighted image was used for segmentation purposes because it shows contrast between the skin, subcutaneous adipose tissue, and muscle. Standard sodium solutions incorporated in the FOV are necessary for calibrating tissue sodium signal intensity to standard sodium concentrations in the physiologic range (10 to 40 mmol/L aqueous NaCl). R: right; L: left; A: anterior; P: posterior.
Sodium imaging was performed using a 3D gradient-echo sequence (TR/TE=130/0.99 ms, field of view (FOV)=192×192 mm2, acquired matrix size=64×64; slice thickness=30 mm, number of signal averages=4) as previously reported9. The extremity was centered over four standard sodium solutions (aqueous NaCl in the physiologic range of peripheral tissue sodium content 10–40 mmol/L) embedded in a platform in the coil on which the extremity rested for 10 minutes prior to acquisition. The sodium image acquisition time for each extremity was approximately 15 minutes.
Without repositioning the participant, the body coil was used for proton radiofrequency transmission and reception. The CSE water-fat method was performed to separate the proton signal from fat and water species in the same acquisition, and provided fat-weighted and water-weighted images in a FOV identical to the sodium image (TR=200 ms, TE1=1.15 ms, TE2=2.30 ms, matrix size=192×192, in-plane spatial resolution=1×1 mm2, 6 slices each 5 mm thick). The CSE image acquisition time for each extremity was approximately 4 minutes. The total multi-nuclear MRI exam of the upper- and lower-extremities including re-positioning lasted approximately 50 minutes.
Analysis
BMI (kg/m2) was calculated from height and weight measurements. Perometry measurements of limb volume (mL) and length (mm) were averaged between right and left sides of the upper or lower extremities, given the bilateral nature of lipedema. Perometry measurements of circumference (cm) were averaged between right and left sides for each location of interest: calf, ankle, forearm, and upper-arm.
Standardized tissue sodium content (TSC) maps were calculated in the arm and leg. First, the mean signal intensity was measured in the standard sodium solutions, and a linear fit (intercept unconstrained) was applied. Model parameters were used to calculate voxel-wise tissue sodium content (mmol/L) from magnitude signal intensity (arbitrary units). TSC maps were interpolated to match the in-plane spatial resolution (1×1 mm2) of CSE images acquired in the identical FOV.
Next, mean TSC was calculated in three regions of interest (ROIs) for each limb: skin, SAT, and muscle. ROIs were manually drawn on the central slice of the CSE water-weighted image and overlaid on the sodium map. The muscle ROI contained all muscle groups in the slice, which consisted of the medial and lateral head of the gastrocnemius, soleus, tibialis posterior, tibialis anterior, and peroneus longus muscles; the total muscle ROI excluded the bone and macrovessels visible on the CSE water-weighted image. The skin region was limited to the posterior semi-perimeter of the skin because this region is subject to less partial volume effects than the anterior extremity.
CSE images were analyzed to measure fat/water ratio and limb circumference using a custom automated segmentation routine (MATLAB, MathWorks, R2018a). First, a k-means clustering algorithm (hard clustering at k=2) was applied to the fat-weighted and water-weighted images to classify voxels as either fat or water (binary). The fat cluster (primarily SAT) was preserved from the fat-weighted image excluding the bone marrow, and the water cluster (primarily muscle) preserved from the water-weighted image. SAT and muscle cross-sectional areas (mm2) were calculated from the sum of voxels in each cluster (voxel dimension=1×1 mm2). For each image, one value was calculated as the total fat/water cross-sectional area (ratio), and reflects the amount of fat relative to the amount of muscle in this portion of the limb. Limb circumference derived from CSE MRI was measured as the sum of voxels populating the perimeter of the water-weighted images of the calf or forearm, and converted to units of cm.
Statistical procedures
Statistical analyses were performed in MATLAB (MathWorks, R2018a, Mathworks, Natick, MA). Descriptive statistics and tests for normality were performed, and all continuous variables were found to be normally distributed. Values are reported as mean±standard deviation for continuous variables or frequencies for categorical variables unless otherwise noted. For all hypothesis testing, significance was defined as two-sided p ≤ 0.05.
Our first statistical objective was to ensure control and lipedema groups were matched for age, BMI, sex, and race using an unpaired Student’s t-test. Physical measurements of blood pressure, limb volume, and limb circumference were also compared using a Student’s t-test.
Our second statistical objective was to test the hypothesis that (i) TSC and (ii) fat/water ratio are different in lower, but not upper, extremities of lipedema compared to control participants. To evaluate this hypothesis, an unpaired Student’s t-test was applied to compare TSC and fat/water ratio in the forearm and calf between lipedema and control participants, and comparisons were performed separately for lower and upper extremities.
Next, in exploratory analyses we evaluated whether the lower vs. upper extremity metrics of tissue sodium or fat/water were different in each group. A paired Student’s t-test was applied to TSC and fat/water ratio in the calf vs. forearm, and comparisons were performed separately for lipedema and control participants. Next, the ratio of calf/forearm TSC and fat/water were computed for each individual, as well as the difference in calf-forearm TSC and fat/water; an unpaired Student’s t-test was applied to determine if these metrics differ between lipedema and control groups.
We also evaluated whether physical measurements of systolic BP, calf circumference, and lower extremity volume were related to TSC and fat/water ratio using a Spearman’s rank correlation test. As these measurements could influence the MRI comparisons performed above, we evaluated the effect of these measurements as covariates with the group effect on MRI metrics using analysis of covariance (ANCOVA, Supplemental Methods).
Finally, we evaluated potential relationships between study measures of calf skin TSC, calf fat/water ratio, or lower extremity volume and established clinical indicators of lipedema severity. Here, the Spearman’s rank correlation coefficient was calculated between study measures and pain (horizontal Visual Analog Scale, VAS from 0–100), or lipedema stage (1, 2, or 3).
Results
Participant characteristics
Participants with lipedema (n=15) had a mean age=43.2±10.0 years and mean BMI=30.3±4.4 kg/m2. Control participants (n=14) had a mean age=42.8±13.2 years and mean BMI=28.8 ± 4.4 kg/m2 (Table 1). All participants were female as lipedema primarily affects women. These groups were matched for age (p=0.92) and BMI (p=0.39), and all participants were Caucasian.
Table 1.
Summary of demographic information and characteristics of participants with lipedema and controls matched for age, BMI, sex, and race.
| Controls | Lipedema | |
|---|---|---|
| Number of participants | 14 | 15 |
| Number of participants with lipedema stage (1, 2, or 3) | (3, 8, 4) | |
| Number of participants with lipedema type (1, 2, or 3) | (1, 5, 9) | |
| Age (years)† | 42.8 ± 13.2 | 43.2 ± 10.0 |
| BMI (kg/m2)† | 28.8 ± 4.4 | 30.3 ± 4.4 |
| Systolic blood pressure (mmHg)* | 114.4 ± 10.0 | 124.0 ± 11.4 |
| Diastolic blood pressure (mmHg) | 74.2 ± 7.3 | 73.6 ± 11.0 |
| Pain (VAS 0–100)* | 3.6 ± 6.7 | 44 ± 24 |
| Easy bruising (% yes) | 14.3 | 93.3 |
| Swelling in affected areas in hot weather (% yes) | 7.1 | 73.3 |
| Fatigue (% yes) | 21.4 | 66.7 |
| Altered gait (% yes) | 0.0 | 60.0 |
| Loose-feeling skin and fat around the knee joint (% yes) | 0.0 | 53.3 |
| Skin in affected areas are cold to the touch (% yes) | 7.1 | 53.3 |
| Hypermobility of the joints (% yes) | 14.3 | 46.7 |
| Use of sleep aids (% yes) | 7.1 | 20.0 |
| Use of diuretics (% yes) | 0.0 | 20.0 |
| Use of pain medications (% yes) | 0.0 | 20.0 |
p>0.05 indicates groups are matched for this parameter.
p ≤ 0.05 indicates a significant group difference for this parameter.
Participants with lipedema had stage 1 (n=3), stage 2 (n=8), or stage 3 (n=4). The majority had involvement of the lower extremities through the calves (type 3: n=9/15). All participants with lipedema had involvement of the upper extremities (n=15/15). Participants with lipedema reported significantly greater pain experienced daily in their lower extremities (lipedema pain=44±24 VAS, control pain=3.6±6.7 VAS, p<0.001). The most frequent symptoms reported by patients with lipedema vs. controls were easy bruising (93.3% vs. 14.3%), swelling in affected areas in hot weather (73.3% vs. 7.1%), and fatigue (66.7% vs. 21.4%). When asked how often fatigue was experienced, 80% of patients with fatigue reported experiencing fatigue daily compared to 33% of controls with fatigue. Recorded characteristics of lipedema and control participants are summarized in Table 1.
Physical measurements
Systolic BP was significantly higher in participants with lipedema (123.9±11.4 mmHg) compared to controls (114.4±10.0 mmHg, p=0.02). Perometry revealed significantly greater lower extremity volume (11,042±2,865 vs. 8,454±1,479 mL, p=0.01) but not upper extremity volume (4,617±393 vs. 4,393±588 mL, p=0.24) in participants with lipedema compared to controls (Table 2). Ankle circumference from perometry (27.1±4.6 vs. 23.0±2.7 cm, p=0.01) and calf circumference from MRI (43.6±3.6 vs. 39.1±4.2 cm, p=0.01) were also significantly larger in participants with lipedema. Upper extremity circumferences were not statistically different between groups by either modality.
Table 2.
Study measurements (mean ± standard deviation) in the arms and legs of participants with lipedema compared to controls.
| Modality | Location | Physiologic Measure | Control | Lipedema | *p-value | Cohen’s d |
|---|---|---|---|---|---|---|
| Sodium MRI | calf | skin sodium (mmol/L) | 14.4±2.2 | 16.3±2.6 | *0.04 | 0.75 |
| calf | SAT sodium (mmol/L) | 13.1±3.1 | 12.6±3.2 | 0.65 | 0.18 | |
| calf | muscle sodium (mmol/L) | 18.3±1.7 | 20.3±3.0 | *0.03 | 0.79 | |
| forearm | skin sodium (mmol/L) | 12.0±2.3 | 13.4±3.3 | 0.20 | 0.50 | |
| forearm | SAT sodium (mmol/L) | 12.3±2.9 | 12.3±2.4 | 0.95 | 0.02 | |
| forearm | muscle sodium (mmol/L) | 16.8±1.2 | 17.5±2.1 | 0.29 | 0.41 | |
| CSE MRI | calf | fat/water (ratio) | 0.56±0.21 | 1.03±0.37 | *<0.001 | 1.23 |
| calf | SAT area (mm2) | 3975±1751 | 6848±2298 | *<0.001 | 1.15 | |
| calf | muscle area (mm2) | 6908±1006 | 6676±1043 | 0.55 | 0.23 | |
| calf | circumference (cm) | 39.1±4.2 | 43.6±3.6 | *0.01 | 1.0 | |
| forearm | fat/water (ratio) | 0.48±0.24 | 0.65±0.24 | 0.07 | 0.68 | |
| forearm | SAT area (mm2) | 1310±832 | 1630±564 | 0.24 | 0.45 | |
| forearm | muscle area (mm2) | 2957±719 | 2750±527 | 0.39 | 0.33 | |
| forearm | circumference (cm) | 25.0±4.1 | 25.7±1.9 | 0.55 | 0.23 | |
| Perometry | lower extremity | volume (mL) | 8454±1479 | 11042±2865 | *0.01 | 0.99 |
| lower extremity | length (mm) | 692±42 | 704±46 | 0.46 | 0.28 | |
| calf | circumference (cm) | 38.5±4.0 | 41.8±6.0 | 0.10 | 0.61 | |
| ankle | circumference (cm) | 23.0±2.7 | 27.1±4.6 | *0.01 | 0.97 | |
| upper extremity | volume (mL) | 4393±588 | 4617±393 | 0.24 | 0.45 | |
| upper extremity | length (mm) | 672±37 | 680±17 | 0.50 | 0.27 | |
| forearm | circumference (cm) | 25.6±5.3 | 25.5±2.2 | 0.96 | 0.02 | |
| upper-arm | circumference (cm) | 32.8±4.1 | 35.2±3.4 | 0.10 | 0.62 | |
SAT: subcutaneous adipose tissue
two-sided significance level p ≤ 0.05
MRI of tissue sodium and fat content
Sodium imaging revealed significantly higher calf TSC in participants with lipedema compared to controls in skin (16.3±2.6 vs. 14.4±2.2 mmol/L, p=0.04) and muscle (20.3±3.0 vs. 18.3±1.7 mmol/L, p=0.03, Figure 2). In the forearm, a moderate effect size was measured between participants with lipedema and controls in skin TSC (13.4±3.3 vs. 12.0±2.3 mmol/L, p=0.2, Cohen’s d=0.50, Table 2).
Figure 2.
MRI metrics in controls (blue) and participants with lipedema (red) in the upper and lower extremities. Tissue sodium content in the a) calf and b) forearm are displayed in the skin, subcutaneous adipose tissue (SAT) and muscle. c) Fat/water ratio was measured in the calf and forearm. Boxplots represent the group median (central bar), upper and lower quartiles (upper and lower bars), and minimum and maximum datapoints (whiskers). *Student’s t-test, two-sided p<0.05; † Paired Student’s t-test, two-sided p<0.05.
CSE MRI revealed significantly higher fat/water ratio in the calf of participants with lipedema compared to controls (1.03±0.37 vs. 0.56±0.21 ratio, p<0.001, Figure 2). SAT cross-sectional area in the calf was significantly greater in participants with lipedema compared to controls (6,848±2,298 vs. 3,975±1,751 mm2, p<0.001), whereas muscle area did not differ between groups. There was a trend for higher fat/water ratio in the forearm of participants with lipedema compared to controls (0.65±0.24 vs. 0.48±0.24 ratio, p=0.07, effect size Cohen’s d=0.68). Forearm cross-sectional area of SAT and muscle were not statistically different between groups (Table 2). These findings corroborate the sodium findings and are consistent with SAT accumulating preferentially in the lower extremities of the participants with lipedema.
Paired comparisons reveal that TSC is significantly higher in the calf skin compared to forearm skin in both lipedema (p<0.001) and control (p<0.001) groups, and in calf muscle compared to forearm muscle in lipedema (p<0.001) and control (p=0.02) groups. TSC in the SAT is not significantly different in the calf compared to forearm in either group. Furthermore, the ratio of calf/forearm skin sodium in lipedema vs. controls is 1.24±0.21 vs. 1.22±0.18 (p=0.78, Cohen’s d=0.11). The difference of calf-forearm skin sodium in lipedema vs. controls is 2.79 ± 2.44 vs. 2.45 ± 1.8 mmol/L (p=0.68, Cohen ‘ s d=0.16). Fat/water ratio is significantly higher in the calf compared to forearm in participants with lipedema (p<0.001) but not in controls (p=0.09); the ratio in calf/forearm is significantly higher in lipedema vs. controls 1.66±0.44 vs. 1.31±0.40 (p=0.04, Cohen’s d=0.76), as well as the difference calf-forearm in lipedema vs. controls 0.42±0.30 vs. 0.08±0.16 (p<0.01, Cohen’s d=1.16). These findings provide evidence that fat deposition is disproportionate in lower vs. upper extremities of participants with lipedema, and distinct from females without lipedema matched for BMI.
Multivariate analyses
The potential effect of systolic BP, lower extremity volume, and calf circumference on imaging metrics of calf skin TSC and fat/water ratio were evaluated. Significant positive correlations were found between systolic BP and calf fat/water ratio among controls (p=0.02), and between calf circumference and calf fat/water ratio among controls (p<0.001) or participants with lipedema (p=0.02, Supplemental Table 1). When accounting for covariates of systolic BP or calf circumference, the group effect on calf fat/water remained significant (Supplemental Tables 2,3).
Relationships between imaging and indicators of disease severity
Among participants with lipedema, there was a significant positive correlation between pain and calf skin TSC (Spearman’s rho=0.55, p=0.03). Calf skin TSC was also positively correlated with lipedema stage (Spearman’s rho=0.82, p<0.001). These findings are summarized in Figure 3 and Table 3.
Figure 3.
a) Calf skin sodium content has a significant, positive relationship with pain (visual analogue scale, VAS). The linear root-mean-squared best-fit line, Spearman’s rho, and p-value are shown. b) Calf skin sodium increases in higher stages of lipedema, and is significantly correlated with stage.
Table 3.
Relationships between study measurements and clinical indicators of disease severity among patients with lipedema.
| Clinical measure | Objective metric | Spearman’s rho | *p-value |
|---|---|---|---|
| Pain (VAS, 0–100) | Calf skin sodium | 0.55 | *0.03* |
| Calf fat/water | −0.064 | 0.82 | |
| Lower extremity volume | 0.15 | 0.59 | |
| Stage (1, 2, 3) | Calf skin sodium | 0.82 | *<0.001 |
| Calf fat/water | −0.19 | 0.50 | |
| Lower extremity volume | 0.36 | 0.19 |
VAS: visual analogue scale
p ≤ 0.05 indicates statistical significance of the correlation
Case example
Representative images of a participant with lipedema and a female without lipedema are displayed in Figure 4. Greater SAT is observed along the legs in whole-body fat-weighted CSE MRI, and in the cross-section of the calf. Elevated TSC is apparent in the calf of the participant with lipedema (see quantitative MRI measurements in figure legend). The forearms appear to have similar SAT and muscle areas, as well as TSC. These features demonstrate the group-level results in representative participants, and specifically that adipose tissue and TSC deposition occurs to a greater degree in the lower extremities compared to upper extremities in participants with lipedema.
Figure 4.
A female without lipedema (control, age=45 years, BMI=38.5 kg/m2) is compared to a female with lipedema (age=38 years, BMI=36.9 kg/m2) on MRI of the a) whole-body, b) forearm, and c) calf. In panels (b-c), images are displayed of CSE water-weighted MRI (top), CSE fat-weighted MRI (middle), and tissue sodium content (bottom). The patient has stage 3 lipedema and experiences symptoms of easy bruising, altered gait, depression, and daily fatigue. MRI reveals a similar circumference of the forearm and calf between participants (forearm circumference lipedema vs. control: 30.4 vs. 29.9 cm; calf circumference lipedema vs. control: 48.2 vs. 45.9 cm). Greater SAT area can be observed in the calf (arrowhead, white pixels surrounding the dark muscle) of the participant with lipedema (9,027 mm2) compared to control (6,715 mm2), contributing to a higher fat/water ratio in lipedema vs. control (1.1 vs. 0.76). Quantitative skin sodium content is elevated in the calf skin (arrow, lipedema vs. control: 17.6 vs. 14.2 mmol/L) and muscle (circle in a region representative of the total muscle sodium content, lipedema vs. control: 23.4 vs. 16.5 mmol/L), while tissue sodium is not apparently different in the forearm. Note the appearance of high sodium content (dark red) in central blood vessels of the arm and leg in both participants; large blood vessels visible on the water-weighted image were excluded from the analysis.
Discussion
MRI observes lower extremity lipedema involvement
Tissue sodium and fat content observable by MRI are significantly elevated in the lower extremities, but not upper extremities, of patients with lipedema compared to BMI-matched controls. Results are consistent with previous findings of elevated skin sodium and fat/water ratio in the calf of participants with lipedema compared to matched controls9, and demonstrate reproducibility in an independent cohort. For participants with lipedema and controls, TSC is significantly higher in the lower vs. upper extremities; only lipedema participants display significantly higher fat/water in the lower vs. upper extremities. Perometry also shows significantly larger volume of lower extremities, but not upper extremities, of participants with lipedema. MRI and perometry observations corroborate bioimpedance spectroscopy measurements which indicate the lower extremities are involved in patients with lipedema compared to their upper extremities, or lower extremities of BMI-matched controls or patients with the adipose tissue disorder Dercum’s disease11.
Higher fat/water measurement in the lower extremities of patients with lipedema reflects larger SAT cross-sectional area in the calf, while muscle size was similar to controls. This does not rule out possible disease effects on muscle function and energy expenditure in lipedema12. Adipose tissue hypertrophy was recently observed histologically13,14 in patients with lipedema, and is consistent with thick SAT observed on MRI9 and ultrasound15,16 throughout the legs. This may be a protective response to a subclinical vascular disease mechanism in lipedema related to sodium dysregulation. For instance, elevated tissue sodium is an emerging hallmark of hypertension17, systemic sclerosis18, infection19, diabetes mellitus20, and chronic kidney disease21. However patients with lipedema are reported to have a lower risk for diabetes and hypertension compared to an obese population10, despite having elevated tissue sodium levels. This suggests an alternative mechanism of sodium deposition, and potentially protective effects of SAT deposition against metabolic syndrome in patients with lipedema.
In the upper extremities, we report moderate effect sizes for elevated skin TSC (Cohen’s d=0.5) and fat/water ratio (Cohen’s d=0.68) content in participants with lipedema with clinically confirmed arm involvement, compared to controls. However, imaging measurements were limited to the forearm, which was not the region of greatest girth in all participants. Clinical observations of lipedema report involvement of different portions of the limbs described by lipedema type6, while the hands and feet are spared. Further cross-sectional studies in participants with specific types of lipedema involvement could aid our understanding of local pathophysiology.
Imaging metrics relate to clinical indicators of disease severity
Significant relationships were observed between calf skin sodium and (i) lipedema stage and (ii) lower extremity pain. These relationships are consistent with symptom advancement in higher stages of lipedema6, and indicate a dysregulation of tissue sodium possibly associated with vascular clearance insufficiency or inflammation. The venous and lymphatic systems are responsible for maintaining plasma concentrations of sodium in the capillary bed as well as the interstitium to maintain sodium homeostasis22. If these systems are impaired, accumulation of sodium over time may ensue as disease advances. Venolymphatic circulation is also resisted by hydrostatic pressure and gravity. Our finding of elevated lower extremity sodium is consistent with gravity-dependent clearance insufficiency. The upper extremities, not subject to these pressures, would not be expected to bear the same edema. Other forces impacting extravasation may be related to a fibrotic extracellular matrix13, permeable vessels8, and impaired venoartieriolar reflex23 reported in patients with lipedema.
Elevated tissue sodium correlates with lower extremity pain in participants with lipedema. This finding is clinically important because i) these are distinguishing characteristics of lipedema from common obesity, and ii) this association may motivate new treatment interventions for pain and in turn TSC may be used to evaluate therapeutic response.
Elevated tissue sodium and pain could be indicators of inflammation in lipedema pathophysiology. For instance, tissue sodium is altered in other diseases of pain and inflammation. Sodium is elevated in soft tissues of patients with scleroderma18 and after acute injury28; tissue sodium is reduced in load-bearing articular cartilage in osteoarthritis30 and intervertebral discs with degenerative disc disease29. While the specific source of pain in lipedema remains unknown, pain on palpation26, joint pain and nociceptive pain10, central sensitization27, and autonomic peripheral neuropathies1 are suggested and should be investigated further using sodium MRI in well-characterized cohorts with lipedema.
Limitations and technology availability
Due to the uncommon diagnosis of lipedema, the sample size was limited to 29 participants. However, these participants were well-characterized with questionnaires, biophysical bedside measurements, and proton and sodium MRI. TSC was elevated in the skin and muscle of participants with lipedema in this study. Muscle sodium was primarily elevated in participants with more advanced stages 2–3 lipedema, while skin sodium progressively increased with stage. The involved regions are heterogeneous compared to prior findings of skin and SAT sodium elevation, and likely reflects the heterogeneity of cases with lipedema. Sodium MRI in the lower extremities has been demonstrated to have a high degree of repeatability31, consistent with the experience at our institution (Supplemental Figure 1). Importantly, known effects of age and sex on tissue sodium variability31–33 were controlled by matching study groups for these parameters. The scope of this study did not permit complete characterization of potential confounds due to salt-sensitive blood pressure, sodium intake, or physical activity level. Future work with larger cohorts may allow additional covariates to be included with sufficient statistical rigor.
Extending these methods to clinical practice presents some challenges. Our experimental implementation of sodium MRI uses a radiofrequency volume coil with an inner circumference of 50 cm, which limits our assessment to the distal extremities and to participants with a compatible calf circumference. Measurement of sodium in the skin using a volume coil at 3.0T is subject to partial volume effects without sub-millimeter voxel resolution, which underestimates true skin sodium. Thus, significant differences in skin sodium may become more robust with technical improvements to sodium MRI. Sodium MRI technology is commercially available at 3.0T clinical field strength, although such technologies are not readily available in most medical centers. Until further development of multi-nuclear imaging technology, sodium MRI may have the greatest impact as a valuable research tool; it provides spatially resolved measurement of tissue sodium content associated with disease severity and pain, and could potentially inform molecular mechanisms of disease and therapeutic response in intervention trials.
CSE MRI of tissue water and fat can be used to assess body composition and is a widely accessible modality with potential to provide objective diagnostics for lipedema. Lower extremity fat/water content and volume measurements provide robust distinctions between patients with moderate-to-advanced lipedema and BMI-matched controls. Measurements were acquired using MRI and perometry, which benefit from high spatial resolution and reduced user dependence. The MRI metric fat/water (ratio) is a proxy for fat-fraction often measured by CSE MRI using longer echo trains to acquire voxel-wise quantitation of proton density fat-fraction, which could be investigated in future studies of lipedema. Fat/water ratio analysis could be improved using soft clustering to increase sensitivity to voxels that have a mix of fat and non-fat signal, although the relatively high spatial-resolution (voxel dimension=1×1 mm2) of CSE MRI minimizes partial volume effects. Other available modalities for assessing limb volume include DEXA, bioimpedance spectroscopy, and 3D surface-rendering technologies, which were not applied here but may be useful to inform diagnosis of lipedema.
Conclusion
Lipedema remains an understudied adipose tissue disorder commonly misdiagnosed as obesity, and lacks an objective characterization of disease pathophysiology. In this study we investigated MRI metrics of tissue sodium and fat content in the upper and lower extremities of patients with lipedema compared to controls. We provide evidence that in the lower extremities tissue sodium is elevated in the skin and muscle of patients with lipedema compared to controls; this finding reproduces a prior report in an independent patient group. In the upper extremities, tissue sodium and fat content were not significantly different and moderate effect sizes were observed. Elevated lower extremity fat content between participants with lipedema vs. controls remained significant when accounting for the effect of blood pressure or calf circumference on fat quantification. Furthermore, fat content in the lower vs. upper extremities was significantly different and thus disproportionate in participants with lipedema, distinct from controls. Together, tissue sodium and fat content could inform objective diagnostic criteria to distinguish lipedema from obesity, and can be achieved noninvasively at clinical strength MRI. Furthermore, tissue sodium content increases with disease severity and pain, and future studies should investigate the role of tissue sodium in lipedema pathophysiology.
Supplementary Material
What is already known about this subject?
Lipedema is a disorder that almost exclusively affects females characterized by disproportionate adipose deposition in the lower extremities, and can be challenging to distinguish from obesity.
Noninvasive MRI previously demonstrated elevated tissue sodium and fat content in the legs of patients with lipedema, compared to females without lipedema matched for body-mass-index.
Upper extremity involvement is also reported in patients with lipedema, yet objective measurement of disease characteristics in upper extremities is lacking.
What are the new findings in your manuscript?
Tissue sodium and fat content are not significantly different in the upper extremities of patients with lipedema compared to controls matched for sex, race, age, and body-mass-index.
In the lower extremities, this work reproduces the finding of significantly elevated skin sodium and subcutaneous fat content in an independent group of patients with lipedema compared to controls.
Tissue sodium content is correlated with clinical metrics of pain and disease stage in patients with lipedema.
How might your results change the direction of research or the focus of clinical practice?
MRI measurement of tissue sodium and fat content have potential to aid in differential diagnosis of lipedema from obesity.
Tissue sodium content should be investigated further to better understand mechanisms of pain and disease severity in lipedema.
Acknowledgments
Imaging experiments were performed in the Vanderbilt Human Imaging Core, using research resources supported by the National Institutes of Health (NIH) grant 1S10OD021771-01 and the Institutional National Research Service Award (NRSA) T32 EB001628. We are grateful to Charles Nockowski, Ed Mojahed, Christopher Thompson, Leslie McIntosh, and Clair Jones for experimental support. Recruitment through www.ResearchMatch.org is supported by the National Center for Advancing Translational Sciences (NCATS) Clinical Translational Science Award (CTSA) Program, award number 5UL1TR002243-03. Funding was provided by the Lipedema Foundation (LF) Postdoctoral Research Fellowship, LF Collaborative Grant #12, and the NIH/NINR 1R01NR015079. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Grant funding:
NIH/NINR 1R01NR015079, NIH 1S10OD021771-01, National Center for Advancing Translational Sciences (NCATS) Clinical Translational Science Award (CTSA) Program award number 5UL1TR002243-03, Institutional National Research Service Award (NRSA) T32 EB001628
Lipedema Foundation Postdoctoral Research Fellowship
Lipedema Foundation Collaborative Grant #12
Declaration of conflict of interest: M.J.D. receives research related support from Philips North America and is the CEO of biosight, LLC which provides healthcare technology consulting services.
References
- 1.Herbst KL Rare adipose disorders (RADs) masquerading as obesity. Acta Pharmacologica Sinica 33, 155–172, doi: 10.1038/aps.2011.153 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Fife CE, Maus E. a. & Carter MJ Lipedema: a frequently misdiagnosed and misunderstood fatty deposition syndrome. Advances in skin & wound care 23, 81–92; quiz 93–94, doi: 10.1097/01.ASW.0000363503.92360.91 (2010). [DOI] [PubMed] [Google Scholar]
- 3.Buso G et al. Lipedema: A Call to Action! Obesity (Silver Spring) 27, 1567–1576, doi: 10.1002/oby.22597 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Al-Ghadban S, Herbst KL, Bunnell BA Lipedema: A Painful Adipose Tissue Disorder [Online First], IntechOpen, DOI: 10.5772/intechopen.88632. Available from: https://www.intechopen.com/online-first/lipedema-a-painful-adipose-tissue-disorder. (2019). [DOI] [Google Scholar]
- 5.Pouwels S, Huisman S, Smelt HJM, Said M & Smulders JF Lipoedema in patients after bariatric surgery: report of two cases and review of literature. Clin Obes 8, 147–150, doi: 10.1111/cob.12239 (2018). [DOI] [PubMed] [Google Scholar]
- 6.Herbst KL, Mirkovskaya L, Bharhagava A, Chava Y & Te CHT Lipedema fat and signs and symptoms of illness, increase with advancing stage. Arch Med. 7, 1–8 (2015). [Google Scholar]
- 7.Duewell S et al. Swollen lower extremity: role of MR imaging. Radiology 184, 227–231 (1992). [DOI] [PubMed] [Google Scholar]
- 8.Lohrmann C, Foeldi E & Langer M MR imaging of the lymphatic system in patients with lipedema and lipo-lymphedema. Microvasc Res 77, 335–339, doi: 10.1016/j.mvr.2009.01.005 (2009). [DOI] [PubMed] [Google Scholar]
- 9.Crescenzi R et al. Tissue Sodium Content is Elevated in the Skin and Subcutaneous Adipose Tissue in Women with Lipedema. Obesity (Silver Spring) 26, 310–317, doi: 10.1002/oby.22090 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Torre YS, Wadeea R, Rosas V & Herbst KL Lipedema: friend and foe. Horm Mol Biol Clin Investig 33, doi: 10.1515/hmbci-2017-0076 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Crescenzi R et al. Lipedema and Dercum’s Disease: A New Application of Bioimpedance. Lymphat Res Biol, doi: 10.1089/lrb.2019.0011 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Herbst KL, Coviello AD, Chang A & Boyle DL Lipomatosis-associated inflammation and excess collagen may contribute to lower relative resting energy expenditure in women with adiposis dolorosa. Int J Obes (Lond) 33, 1031–1038, doi: 10.1038/ijo.2009.119 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Al-Ghadban S et al. Dilated Blood and Lymphatic Microvessels, Angiogenesis, Increased Macrophages, and Adipocyte Hypertrophy in Lipedema Thigh Skin and Fat Tissue. J Obes 2019, 8747461, doi: 10.1155/2019/8747461 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bauer AT et al. Adipose stem cells from lipedema and control adipose tissue respond differently to adipogenic stimulation in vitro. Plast Reconstr Surg, doi: 10.1097/PRS.0000000000005918 (2019). [DOI] [PubMed] [Google Scholar]
- 15.Iker E, Mayfield CK, Gould DJ & Patel KM Characterizing Lower Extremity Lymphedema and Lipedema with Cutaneous Ultrasonography and an Objective Computer-Assisted Measurement of Dermal Echogenicity. Lymphat Res Biol, doi: 10.1089/lrb.2017.0090 (2019). [DOI] [PubMed] [Google Scholar]
- 16.Dimakakos PB et al. MRI and ultrasonographic findings in the investigation of lymphedema and lipedema. Int Surg 82, 411–416 (1997). [PubMed] [Google Scholar]
- 17.Kopp C et al. 23Na magnetic resonance imaging-determined tissue sodium in healthy subjects and hypertensive patients. Hypertension 61, 635–640, doi: 10.1161/HYPERTENSIONAHA.111.00566 (2013). [DOI] [PubMed] [Google Scholar]
- 18.Kopp C et al. Na+ deposition in the fibrotic skin of systemic sclerosis patients detected by 23Na-magnetic resonance imaging. Rheumatology (Oxford), doi: 10.1093/rheumatology/kew371 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Jantsch J et al. Cutaneous Na+ storage strengthens the antimicrobial barrier function of the skin and boosts macrophage-driven host defense. Cell Metab 21, 493–501, doi: 10.1016/j.cmet.2015.02.003 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Deger SM et al. Tissue sodium accumulation and peripheral insulin sensitivity in maintenance hemodialysis patients. J Cachexia Sarcopenia Muscle, doi: 10.1002/jcsm.12179 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Schneider MP et al. Skin Sodium Concentration Correlates with Left Ventricular Hypertrophy in CKD. J Am Soc Nephrol, doi: 10.1681/ASN.2016060662 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Swartz M. a. The physiology of the lymphatic system. [Review] [190 refs]. Advanced Drug Delivery Reviews 50, 3–20 (2001). [DOI] [PubMed] [Google Scholar]
- 23.Burnett K, Herbst KL. The Venoarteriolar Reflex is Abnormal in the Legs of Women with Stage 2 and 3 Lipedema Compared to Controls. 32rd Annual Congress of the Americal College of Phlebology Abstract presented November 10, 2019 (2019). [Google Scholar]
- 24.Wang C, Witschey W, Elliott MA, Borthakur A & Reddy R Measurement of intervertebral disc pressure with T 1rho MRI. Magn Reson Med 64, 1721–1727, doi: 10.1002/mrm.22560 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Borthakur A et al. Sodium and T1rho MRI for molecular and diagnostic imaging of articular cartilage. NMR Biomed 19, 781–821, doi: 10.1002/nbm.1102 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Halk AB & Damstra RJ First Dutch guidelines on lipedema using the international classification of functioning, disability and health. Phlebology 32, 152–159, doi: 10.1177/0268355516639421 (2017). [DOI] [PubMed] [Google Scholar]
- 27.Nijs J et al. Pain following cancer treatment: Guidelines for the clinical classification of predominant neuropathic, nociceptive and central sensitization pain. Acta Oncol 55, 659–663, doi: 10.3109/0284186X.2016.1167958 (2016). [DOI] [PubMed] [Google Scholar]
- 28.Dahlmann A et al. Quantitative assessment of muscle injury by (23)Na magnetic resonance imaging. Springerplus 5, 661, doi: 10.1186/s40064-016-2193-6 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wang C et al. Validation of sodium magnetic resonance imaging of intervertebral disc. Spine (Phila Pa 1976) 35, 505–510, doi: 10.1097/BRS.0b013e3181b32d3b (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Newbould RD et al. Reproducibility of sodium MRI measures of articular cartilage of the knee in osteoarthritis. Osteoarthritis Cartilage 20, 29–35, doi: 10.1016/j.joca.2011.10.007 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Dyke JP et al. Reliability and agreement of sodium ((23)Na) MRI in calf muscle and skin of healthy subjects from the US. Clin Imaging 52, 100–105, doi: 10.1016/j.clinimag.2018.07.010 (2018). [DOI] [PubMed] [Google Scholar]
- 32.Kopp C et al. 23Na magnetic resonance imaging-determined tissue sodium in healthy subjects and hypertensive patients. Hypertension 61, 635–640, doi: 10.1161/HYPERTENSIONAHA.111.00566 (2013). [DOI] [PubMed] [Google Scholar]
- 33.Wang P et al. Sex differences in sodium deposition in human muscle and skin. Magn Reson Imaging 36, 93–97, doi: 10.1016/j.mri.2016.10.023 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.