Associations of myosteatosis with disc degeneration: A 3T magnetic resonance imaging study in individuals with impaired glycaemia

Abstract

Background

Intervertebral disc degeneration (IVDD) may be linked to dysregulations of skeletal muscle glucose metabolism and fatty alterations of muscle composition (Myosteatosis). Our aim was to evaluate the different associations of magnetic resonance imaging (MRI)‐based paravertebral myosteatosis with lumbar disc degeneration in individuals with impaired glucose metabolism and normoglycaemic controls.

Methods

In total, 304 individuals (mean age: 56.3 ± 9.1 years, 53.6% male sex, mean body mass index [BMI]: 27.6 ± 4.7 kg/m²) from a population‐based cohort study who underwent 3‐Tesla whole‐body chemical‐shift‐encoded (six echo times) and T2‐weighted single‐shot‐fast‐spin‐echo MRI were included. Lumbar disc degeneration was assessed at motion segments L1 to L5, categorized according to the Pfirrmann score and defined as Pfirrmann grade > 2 and/or disc bulging/herniation on at least one segment. Fat content of the autochthonous back muscles and the quadratus lumborum muscle was quantified as proton density fat fraction (PDFF_muscle). Logistic regression models adjusted for age, sex, BMI and regular physical activity were calculated to evaluate the association between PDFF_muscle and outcome IVDD.

Results

The overall prevalence of IVDD was 79.6%. There was no significant difference in the prevalence or severity distribution of IVDD between participants with or without impaired glucose metabolism (77.7% vs. 80.7%, P = 0.63 and P = 0.71, respectively). PDFF_muscle was significantly and positively associated with an increased risk for the presence of IVDD in participants with impaired glycaemia when adjusted for age, sex and BMI (PDFF_{autochthonous back muscles}: odds ratio [OR] 2.16, 95% confidence interval [CI] [1.09, 4.3], P = 0.03; PDFF_{quadratus lumborum}: OR 2.01, 95% CI [1.04, 3.85], P = 0.04). After further adjustment for regular physical activity, the results attenuated, albeit approaching statistical significance (PDFF_{autochthonous back muscles}: OR 1.97, 95% CI [0.97, 3.99], P = 0.06; PDFF_{quadratus lumborum}: OR 1.86, 95% CI [0.92, 3.76], P = 0.09). No significant associations were shown in healthy controls (PDFF_{autochthonous back muscles}: OR 0.62, 95% CI [0.34, 1.14], P = 0.13; PDFF_{quadratus lumborum}: OR 1.06, 95% CI [0.6, 1.89], P = 0.83).

Conclusions

Paravertebral myosteatosis is positively associated with intervertebral disc disease in individuals with impaired glucose metabolism, independent of age, sex and BMI. Regular physical activity may confound these associations. Longitudinal studies will help to better understand the pathophysiological role of skeletal muscle in those with concomitant disturbed glucose haemostasis and intervertebral disc disease, as well as possible underlying causal relationships.

Introduction

Intervertebral disc degeneration (IVDD) is a chronic, progressive condition, characterized by a decrease of proteoglycans and water content in the nucleus pulposus. ¹ The consecutive loss of disc height leads to biomechanical alterations through changes of load distribution across the spine. ² Degeneration of the spine is thought to be initiated by IVDD, followed by segmental instability and eventually degeneration of facet joints. ³ This sequence highlights the suspected key role of IVDD as the main driver of early spinal osteoarthritis (OA). Furthermore, IVDD seems to be accompanied by concomitant changes in the adjacent vertebral endplates and fatty alterations of paravertebral skeletal musculature, also known as myosteatosis. ⁴

Myosteatosis represents the qualitative aspect of skeletal muscle deterioration and is considered a sensitive and early marker of skeletal muscle depletion. ⁵ , ⁶ In addition, myosteatosis is associated with type 2 diabetes mellitus (T2DM), impaired mobility and frailty. ⁷ , ⁸ Because IVDD has been shown to be more prevalent in individuals with impaired glucose metabolism (prediabetes and T2DM) and/or obesity, epidemiological and causative relations through systemic mechanisms have been proposed between these entities. ⁹ , ¹⁰ Interestingly, skeletal muscle wasting may constitute a possible link between metabolic disturbances, as in T2DM, and joint degeneration, acting as additional risk factor in the development and progression of OA of the lower extremities. ¹¹ , ¹² However, whether an analogous relationship between myosteatosis and ‘metabolic’ IVDD exists remains unclear. The aim of the present cross‐sectional study was to analyse paravertebral myosteatosis using quantitative chemical‐shift encoding‐based water‐fat magnetic resonance imaging (MRI) and to evaluate the different associations of myosteatosis with IVDD in individuals from the general population with and without comorbid impairment of glucose metabolism. Our hypothesis was that myosteatosis is associated with a significantly increased risk for the presence of IVDD in those with impaired glucose metabolism, compared with normoglycaemic controls.

Materials and methods

Study design and population

The investigation was designed as a cross‐sectional study based on a prospective cohort from the ‘Cooperative Health Research in the Region of Augsburg’ (KORA) as described in detail before. ¹³ In brief, participants for this analysis were derived from the KORA‐FF4 study (June 2013 to September 2014, n = 2279), the second follow‐up study of the KORA S4 cohort (baseline survey 1999–2001, n = 4216), which included a large representative sample of individuals aged between 25 and 74 years from the general population. ¹³ A comprehensive health assessment was performed for all participants. In addition, 400 participants underwent whole‐body MRI according to previously described inclusion and exclusion criteria. ¹³ Median time to MRI examination was 33 days (interquartile range [IQR] 24–45). The study was IRB approved by the ethics committee of the Bavarian Chamber of Physicians, Munich, Germany, and the local institutional review board of the Ludwig‐Maximilians University Munich, Germany. The study complies with the Declaration of Helsinki, including written informed consent of all participants.

Magnetic resonance imaging protocol and imaging analysis

MRI examinations were performed in supine position on a 3‐Tesla Magnetom Skyra (Siemens Healthineers, Erlangen, Germany) using an 18‐channel body surface coil in combination with a table‐mounted spine matrix coil. The complete imaging protocol and technical specificities have been described in detail previously. ¹³

Skeletal muscle segmentation

Skeletal muscle segmentation was performed at level L3 vertebra on axial slices using a T2*‐corrected, multi‐echo 3D‐gradient‐echo Dixon‐based sequence (multi‐echo Dixon) with the following parameters (TR: 8.90 ms; TEs: 1.23, 2.46, 3.69, 4.92, 6.15 and 7.38 ms; flip angle: 4°; slice thickness: 4 mm) (Figure  1 ). The accurate transversal slice position at level L3 vertebra was confirmed by identifying L4 vertebra by the iliac crest tangent sign on coronal images using cross‐reference. ¹⁴

Figure 1.

Paravertebral muscle segmentation at the level of the third lumbar vertebra. Multi‐echo 3D‐gradient‐echo Dixon‐based sequence (TR: 8.90 ms; TEs: 1.23, 2.46, 3.69, 4.92, 6.15 and 7.38 ms; flip angle: 4°; slice thickness: 4 mm).

Muscle compartments of M. quadratus lumborum and the autochthonous back muscles were manually segmented according to standardized, anatomical landmarks using a dedicated software (MITK V2015.5.2, German Cancer Research Center, Heidelberg, Germany). Segmentation was performed by two individual observers, fully blinded to clinicopathological data and IVDD status. Proton density fat fraction (PDFF) maps were automatically calculated as DICOM files using the software MR LiverLab (Version VD13, Siemens Healthineers, Cary, USA). The degree of myosteatosis was determined as mean PDFF indicated in percentage. Hereby, the final measure was defined as average PDFF of both the right and left muscle compartments. Intraobserver and interobserver reproducibility of the segmentation algorithm was assessed in a prior study that yielded overall excellent results for all analysed muscle compartments (intraclass correlation coefficient [ICC] 0.95 to 0.97, and ICC 0.97 to 0.98, respectively). ¹⁴

Intervertebral disc degeneration analysis

Degenerative changes of each intervertebral segment were analysed by a sequence‐adapted Pfirrmann score on a T2‐HASTE sequence (TR: 1000 ms, TE: 91 ms, flip angle: 131°, slice thickness: 5 mm). Therefore, structure, distinction between nucleus and annulus, signal intensity and height of intervertebral disc were analysed for each segment from lumbar vertebrae 1 (L1) to lumbar vertebrae 5 (L5). ¹⁵ Detectible intervertebral disc tissue in the spinal canal was analysed on sagittal reconstructed dual‐echo Dixon images and classified in either disc bulging, not exceeding the endplates of adjacent vertebrae, or disc protrusion, a perimeter of disc tissue exceeding the endplates of adjacent vertebrae.

IVDD was defined as Pfirrmann grade > 2 and/or disc bulging/herniation on at least one lumbar segment. This threshold was chosen because Pfirrmann grades 1 and 2 only show a minimal degree of pathology and level of difference. Further, these changes are considered normal findings in aging. Analysis was performed by two independent readers with 5 and 6 years of experience in musculoskeletal imaging, respectively. Readers were fully blinded to clinicopathological data. To assess intra‐reader agreement of continuous measurements, a total of 40 randomly selected cases were re‐evaluated in a blinded fashion by the primary reader. Inter‐ and intra‐reader agreement analysis of the Pfirrmann score was 85% (Kendall W = 0.93) and 87.5% (Kendall W = 0.92), respectively. Inter‐ and intra‐reader agreement analysis of disc bulging and disc protrusion displayed an inter‐ and intra‐reader agreement of 95% (Kappa = 0.84) for disc bulging and an inter‐ and intra‐reader agreement of 100% (Kappa = 1.00) for disc protrusion. ¹⁶ In case of disagreement between readers for Pfirrmann grading, the ratings of the senior reader were considered as the final value.

Health assessment

All participants underwent standardized interviews and physical examinations at the study centre in order to determine main characteristics and demographics as well as major cardiometabolic risk factors.

Glycaemic status—Oral glucose tolerance test

A 75‐g oral glucose tolerance test (OGTT) was performed for all participants, which had not yet been diagnosed with diabetes mellitus (DM) and/or which did not report regular intake of antihyperglycaemic mediation (according to most recent guidelines) after at least 8‐h overnight fasting. Previously undiagnosed diabetes (≥125 mg/dL fasting and/or ≥200 mg/dL 2‐h glucose load), impaired fasting glucose (IFG; fasting glucose 110–125 mg/dL), impaired glucose tolerance (IGT; 2‐h glucose 140–200 mg/dL) and normoglycaemia (fasting < 110 mg/dL and 2‐h glucose < 140 mg/dL) were defined according to the World Health Organization (WHO) criteria. ¹⁷ For the present analysis, we defined all participants with established or newly diagnosed diabetes, IFG and IGT as having impaired glucose metabolism.

Other covariates

Blood samples were assessed for fasting blood levels of glucose and glycated haemoglobin (HbA1c) as well as blood lipids by standard laboratory methods. ¹³ The body mass index (BMI) was calculated as body weight in kilograms divided by body height squared in square metres, with weight and height both being measured at the study centre. Waist circumference was measured at the smallest abdominal circumference or, in individuals with obesity, in the midpoint of the lowest rib and the upper margin of the iliac crest.

Assessment of physical activity

Physical activity was evaluated by self‐report using standardized questionnaires. Subsequently, a dichotomous variable was calculated with (A) physically active participants (regular physical activity > 1 h/week) or (B) physically inactive participants.

Statistical analysis

Baseline characteristics of the study population are presented as mean and standard deviation (±) or median with IQR for continuous variables and absolute counts with percentages for categorical variables. Differences between groups (e.g., individuals with impaired glucose metabolism and normoglycaemic controls) were assessed using t‐test, Mann–Whitney U‐test or χ ²‐test, as appropriate. To assess the associations between PDFF_muscle as a continuous variable and IVDD as a dichotomous outcome logistic regression, models were fitted and adjusted for age, sex and BMI and, in addition, physical activity, as recommended by the International Society for the Study of the Lumbar Spine (ISSLS) Degenerative Spinal Phenotypes Group. ¹⁸

Continuous measures were standardized (by subtracting the mean and dividing by standard deviation) before modelling. Results are presented as odds ratios (ORs) with corresponding 95% confidence intervals (CIs). Statistical significance was indicated by P‐values < 0.05. Statistical analysis was performed using R V4.0.3 (R Core Team, www.r‐project.org, 2020).

Results

Study population

Among 400 participants who underwent supine whole‐body MRI, 96 individuals (24%) were excluded due to insufficient image quality, incomplete MRI data or missing values in any of the covariates (Figure  2 ). Thus, the study cohort consisted of 304 participants (mean age: 56.3 ± 9.1 years, 53.6% male, mean BMI: 27.6 ± 4.7 kg/m²). Overall, 112 out of the 304 participants (36.8%) had an impaired glucose metabolism (43/112 [14.1%] T2DM and 69/112 [22.7%] prediabetes, thereof 26 with isolated IFG, 32 with isolated IGT and 11 with both IFG and IGT). The remaining 192 (63.2%) participants were normoglycaemic. As compared with normoglycaemic individuals, participants with impaired glucose metabolism were older, were more likely male and had a higher BMI (all P < 0.05). Details of the demographics are provided in Table 1 .

Figure 2.

Participant flow diagram. PDFF, proton density fat fraction.

Table 1.

Demographics and characteristics of the study cohort

	Whole sample	Normal glucose tolerance	Impaired glucose metabolism a	P‐value
	n = 304	192 (63.2%)	112 (36.8%)
Age (years)	56.3 ± 9.1	54.1 ± 8.7	60.1 ± 8.6	<0.001
Male gender	163 (53.6%)	89 (46.4%)	74 (66.1%)	0.001
BMI (kg/m2)	27.6 ± 4.7	26.5 ± 4.4	29.6 ± 4.6	<0.001
Waist circumference (cm)	96.8 ± 13.5	92.7 ± 12.6	104.0 ± 12.1	<0.001
Glucose tolerance
Blood glucose level	103.2 ± 20.4	94.2 ± 7.4	119.1 ± 25.6	<0.001
HbA1c, %	5.6 ± 0.6	5.3 ± 0.3	5.9 ± 0.7	<0.001
Lipid metabolism
Triglycerides (mg/dL)	128.0 ± 83.4	104.9 ± 65.0	167.5 ± 95.9	<0.001
Total cholesterol	217.7 ± 35.8	214.8 ± 34.4	222.7 ± 37.6	0.06
HDL (mg/dL)	63.0 ± 17.9	66.3 ± 18.4	57.5 ± 15.6	<0.001
LDL (mg/dL)	139.1 ± 32.2	137.0 ± 30.9	142.8 ± 34.1	0.13
Physical activity
Exercise regularly >2 h/week	86 (28.3%)	59 (30.7%)	27 (24.1%)	0.12
Exercise regularly >1 h/week	101 (33.2%)	68 (35.4%)	33 (29.5%)
Exercise irregularly ≤1 h/week	48 (15.8%)	24 (12.5%)	24 (21.4%)
Almost no/no physical activity	69 (22.7%)	41 (21.4%)	28 (25.0%)

Note: Data are presented as means (SD) for continuous variables and counts and percentages for categorical variables. Boldface is used to highlight significant P‐values.

Abbreviations: BMI, body mass index; HbA1c, glycated haemoglobin.

^a Known and undiagnosed type 2 diabetes mellitus and prediabetes.

Muscle composition and intervertebral disc degeneration

Colour‐coded paravertebral PDFF_muscle maps are shown in Figure 3 .

Figure 3.

(A) Low paravertebral myosteatosis (proton density fat fraction [PDFF] 4.5%) in a normal weight male participant with normal glucose tolerance and without relevant intervertebral disc degeneration (Pfirrmann grade 2). (B) Marked fatty atrophy of the paravertebral muscles (PDFF 26.1%) in an overweight male participant with type 2 diabetes mellitus and severe intervertebral disc degeneration (Pfirrmann grade 5). Blue colour indicates low fat fractions, and red colour indicates high fat fractions.

Overall, median PDFF_muscle of the autochthonous back muscles in all participants was 16.4% (IQR 11.7–21.5) and median PDFF_muscle of the quadratus lumborum was 6.0% (IQR 4.0–8.0). Median PDFF_muscle was significantly higher in individuals with impaired glucose metabolism as compared with normoglycaemic controls for both the autochthonous back muscles (17.8% [IQR 13.3–25.2] vs. 14.9% [IQR 10.7–20.4]; P < 0.001) and the quadratus lumborum muscle (6.5% [IQR 4.4–9.2] vs. 5.4% [IQR 3.8–7.4]; P = 0.003). A description according to different subgroups of glucose metabolism impairment (i‐IFG, i‐IGT, IFG + IGT) is given in Table S1 . The overall prevalence of IVDD was 79.6%. There was no significant difference in the prevalence of IVDD (77.7% vs. 80.7%; P = 0.63) or in severity of IVDD (P = 0.71) between participants with an impaired glucose metabolism and normoglycaemic participants. Furthermore, no statistical difference in the prevalence of disc bulging and/or herniation was observed between both glycaemic status groups with 50.9% and 47.9%, respectively (P = 0.70) (Table  2 ). Differences in the degree of PDFF_muscle in individuals with and without IVDD are depicted in Figure 4 .

Table 2.

Muscle composition and disc degeneration of the studied cohort

	Whole sample	Normal glucose tolerance	Impaired glucose metabolism a	P‐value
	n = 304	192 (63.2%)	112 (36.8%)
PDFFautochthonous back muscles (%)	16.4 [11.7, 21.5]	14.9 [10.7, 20.4]	17.8 [13.3, 25.2]	<0.001
PDFFquadratus lumborum (%)	6.0 [4.0, 8.0]	5.4 [3.8, 7.4]	6.5 [4.4, 9.2]	0.003
IVDD	242 (79.6%)	155 (80.7%)	87 (77.7%)	0.63
Distribution of Pfirrmann scores				0.71
Pfirrmann score 3 L1/L2–L5/S1	88 (28.9%)	53 (27.6%)	35 (31.2%)
Pfirrmann score 4 L1/L2–L5/S1	82 (27.0%)	53 (27.6%)	29 (25.9%)
Pfirrmann score 5 L1/L2–L5/S1	38 (12.5%)	22 (11.5%)	16 (14.3%)
Disc bulging or herniation level L1/L2 to L5/S1	149 (49.0%)	92 (47.9%)	57 (50.9%)	0.70

Note: Data are presented as median [first quartile, third quartile] for continuous variables and counts and percentages for categorical variables. Boldface is used to highlight significant P‐values.

Abbreviations: IVDD, intervertebral disc degeneration; PDFF, proton density fat fraction.

^a Known and undiagnosed type 2 diabetes mellitus and prediabetes.

Figure 4.

Differences in the degree of paravertebral myosteatosis (PDFF_muscle %) between individuals with impaired glucose metabolism and individuals with normal glucose tolerance. (A) PDFF_{autochthonous back muscles}. (B) PDFF_{quadratus lumborum}. IVDD, intervertebral disc degeneration; PDFF, proton density fat fraction.

Associations of myosteatosis with intervertebral disc degeneration

Results from a logistic regression model adjusted for age, sex and BMI with PDFF_muscle as exposure and IVDD as outcome variables are depicted in Table 3 . PDFF_muscle was significantly and positively associated with IVDD in individuals with impaired glucose metabolism (PDFF_{autochthonous back muscles}: OR 2.16, 95% CI [1.09, 4.3], P = 0.03; PDFF_{quadratus lumborum}: OR 2.01, 95% CI [1.04, 3.85], P = 0.04). In contrast, no significant association was found for PDFF_muscle and IVDD, neither in normoglycaemic individuals (PDFF_{autochthonous back muscles}: OR 0.62, 95% CI [0.34, 1.14], P = 0.13; PDFF_{quadratus lumborum}: OR 1.06, 95% CI [0.6, 1.89], P = 0.83) nor in the whole sample (PDFF_{autochthonous back muscles}: OR 1.18, 95% CI [0.76, 1.83], P = 0.46; PDFF_{quadratus lumborum}: OR 1.44, 95% CI [0.95, 2.19], P = 0.08).

Table 3.

Association of PDFF_muscle with outcome IVDD

IVDD	Whole sample			Normal glucose tolerance			Impaired glucose metabolism
	Odds ratio	95% CI	P‐value	Odds ratio	95% CI	P‐value	Odds ratio	95% CI	P‐value
Model 1
PDFFautochthonous back muscles	1.18	[0.76, 1.83]	0.46	0.62	[0.34, 1.14]	0.13	2.16	[1.09, 4.3]	0.03
PDFFquadratus lumborum	1.44	[0.95, 2.19]	0.08	1.06	[0.6, 1.89]	0.83	2.01	[1.04, 3.85]	0.04
Model 2
PDFFautochthonous back muscles	1.18	[0.76, 1.83]	0.47	0.64	[0.34, 1.18]	0.15	1.97	[0.97, 3.99]	0.06
PDFFquadratus lumborum	1.44	[0.95, 2.2]	0.09	1.05	[0.59, 1.89]	0.86	1.86	[0.92, 3.76]	0.09

Note: Results from a logistic regression model, adjusted for age, sex, body mass index (Model 1), and age, sex, body mass index, physical activity (Model 2), outcome IVDD. Continuous risk factors were standardized before modelling; odds ratios denote changes in risk for IVDD associated with an increase of one standard deviation of the continuous risk factor. Boldface is used to highlight significant results.

Abbreviations: CI, confidence interval; IVDD, intervertebral disc degeneration; PDFF, proton density fat fraction.

After further adjustment for physical activity, the associations attenuated in individuals with impaired glucose, with respective P‐values approaching the level of statistical significance (PDFF_{autochthonous back muscles}: OR 1.97, 95% CI [0.97, 3.99], P = 0.06; PDFF_{quadratus lumborum}: OR 1.86, 95% CI [0.92, 3.76], P = 0.09). Analogously, no significant associations of PDFF_muscle with IVDD were found for the normoglycaemic group or whole sample (Model 2). In additional analyses, we tested the association of regular physical activity with IVDD and found that individuals with impaired glucose metabolism who were physically active had a significantly lower risk for IVDD (OR 0.3, 95% CI [0.1, 0.81], P = 0.02). Further, regular physical activity and PDFF_muscle showed a significant inverse correlation in individuals with impaired glucose metabolism (PDFF_{autochthonous back muscles}: ß −2.77, 95% CI [−5.45, −0.09], P = 0.04; PDFF_{quadratus lumborum}: ß −2.16, 95% CI [−3.58, −0.74], P = 0.003) (Tables S2 and S3 ).

Discussion

In this population‐based cross‐sectional study, quantitative MRI techniques were used to assess the association between paraspinal myosteatosis and IVDD.

In the studied cohort, individuals with impaired glucose metabolism and concomitant myosteatosis showed increased odds for IVDD, independent of age, sex or BMI.

However, the associations were confounded by physical activity after further adjustment.

To the best of our knowledge, this is the first study that investigated the potential associations of MRI‐based paravertebral myosteatosis with degenerative disc disease focused on individuals with impaired glucose metabolism. Previous studies that analysed the relationship between paravertebral muscle and disc degeneration were mainly conducted in symptomatic individuals with spinal stenosis or lower back pain. ¹⁹ , ²⁰ , ²¹ In a retrospective study including 72 community‐based individuals, Teichtahl et al. demonstrated a positive association between lumbar IVDD, Modic changes of the vertebral endplates and a high paraspinal fat content, suggesting that IVDD should be considered a ‘whole‐organ’ pathology, notably affecting bone, intervertebral discs and muscle tissue. However, these associations were only found in severe IVDD (Pfirrmann grade ≥ 4), thus possibly reflecting reactive fatty atrophy of the paravertebral muscles due to spinal degeneration. ⁴ , ²² In our cohort, individuals with impaired glucose metabolism showed significantly higher degrees of PDFF_muscle also including less severe IVDD, as compared with normoglycaemic controls, possibly suggesting a rather metabolic related change of paravertebral muscle composition in those with alteration of glucose metabolism. ²³

Current concepts on the development of osteodegenerative disease in individuals with metabolic disorders suggest a complex interplay of several factors eventually leading to oxidative stress and chronic low‐grade inflammation. ²⁴ Regarding their pathophysiology, myosteatosis and deterioration of glucose metabolism in prediabetes and T2DM are believed to share some important risk factors. Intramyocellular lipids represent ectopic lipid deposits within mature myocytes. By producing proinflammatory mediators, intramyocellular lipids have shown to display a metabolic activity similar to other adipose tissue compartments, for example, visceral adipose tissue. This might further aggravate skeletal muscle insulin resistance. ²⁵ Because proinflammatory mediators are known to not only initiate but also accelerate the process of IVDD, myosteatosis may be linked to IVDD in individuals with disturbed glucose metabolism via chronic inflammatory changes of muscle tissue surrounding the IVDD microenvironment. ²⁶ In a recent study including 24 patients undergoing spine surgery for disc herniation, James et al. analysed muscle tissue and epidural fat and described higher tissue levels of inflammatory cytokines in those with a high fat content of the multifidus muscle. ²⁷

Our results support the suspected role of myosteatosis as potential additional risk factor in the context of metabolic osteodegeneration of the spine. However, it is important to state that the cross‐sectional design of the present study is, by virtue, unable to proof causalities. Despite the phenotypic differences between normoglycaemic individuals and those with impaired glucose metabolism, fatty infiltration could also represent a reactive change of paravertebral muscle composition in individuals with IVDD, independent of their respective glycaemic status.

In addition, other risk factors for metabolic IVDD have also been identified. T2DM disease duration and long‐term elevated blood glucose levels seem associated with severe IVDD and hyperglycaemia may directly affect disc degeneration via microangiopathic changes and the formation of glycation end products in the nucleus pulposus. ²⁸ , ²⁹ Furthermore, comorbid overweight and consecutive biomechanical stress on the axial skeleton, as well as a sedentary lifestyle, may additionally contribute to the development and progression of IVDD, especially in individuals with metabolic syndrome and/or T2DM. ³⁰ , ³¹

Interestingly, in our study cohort, regular physical activity confounded the associations of paravertebral myosteatosis with IVDD in those with impaired glucose metabolism. In additional analyses, regular physical activity and PDFF_muscle showed a significant inverse relationship, and physically active individuals with impaired glucose metabolism had a significantly lower risk for IVDD. In line, myosteatosis is considered a modifiable condition and a positive effect of regular physical exercise on myosteatosis has been described previously. ³² Furthermore, regular physical activity is recommended for patients with T2DM and OA as a main therapeutic option. ³³

Body composition analysis using cross‐sectional imaging has been a growing area in medical research over the past years. Single‐slice analyses of the different body compartments at the level of the third lumbar vertebra (e.g., skeletal muscle, subcutaneous adipose tissue and visceral adipose tissue) using axial MRI or computed tomography have been established as imaging surrogate parameters for overall body composition. ³⁴ , ³⁵ Especially measurements of muscle cross‐sectional area and qualitative or quantitative measurements of intramuscular fat have moved into focus, due to their possible association with morbidity and mortality in various diseases. ⁸ , ³⁶ With regard to spinal degeneration, the majority of studies used qualitative or semi‐quantitative methods, such as the visual Goutallier score or cross‐sectional area to analyse fatty infiltration of skeletal muscle. Of note, semi‐quantitative visual grading systems such as the Goutallier score are known to be prone to significant interobserver and intraobserver variability. ³⁷ A strength of our study is the use of chemical‐shift MRI imaging to analyse paravertebral myosteatosis, a quantitative approach enabling the detection of subtle alterations of muscle composition with overall good reproducibility. ³⁸ , ³⁹

However, the main drawback of our study is the retrospective cross‐sectional study design, therefore requiring confirmation in larger, longitudinal cohort studies. Second, participants with impaired glucose metabolism were therapeutically well controlled. However, this cohort represents a sample of a general southern German population that was randomly selected. In addition, MR‐based results of myosteatosis measurements were not compared with histopathology, which is still considered the current gold standard for quantification of fat content, and sample PDFF measurements on a single axial slice at the level of the L3 vertebra, as performed in this study, may not reflect the exact distribution of lipid storage within the whole. However, previous studies have demonstrated the validity and reproducibility of a standardized, anatomic landmark‐based quantification of skeletal muscle fat via measurement of PDFF, as performed in the present study, and overall high concordance of PDFF measurements with histology. ¹⁴ , ⁴⁰

Conclusions

Paravertebral myosteatosis is positively associated with lumbar disc degeneration in individuals with impaired glucose metabolism, independent of age, sex and BMI. However, regular physical activity seems to represent a relevant confounding factor. Our findings underline that paravertebral myosteatosis may constitute an additional, potentially modifiable risk factor for degenerative disc disease in those with comorbid impaired glucose metabolism. Further longitudinal studies are required to elucidate possible cause–effect relationships.

Conflict of interest statement

None.

Funding

This study was funded by the German Research Foundation (DFG, Bonn, Germany), the German Centre for Cardiovascular Disease Research (DZHK, Berlin, Germany) and the Centre for Diabetes Research (DZD e.V., Neuherberg, Germany). The KORA study was initiated and financed by the Helmholtz Zentrum München – German Research Center for Environmental Health, which is funded by the German Federal Ministry of Education and Research (BMBF) and by the State of Bavaria.

Supporting information

Table S1. Distribution of PDFF and IVDD in the respective subgroups of impaired glucose metabolism

i‐IFG: impaired fasting glucose. I‐IGT: impaired glucose tolerance. IVDD: intervertebral disc degeneration. PDFF: proton‐density fat‐fraction

Table S2. Association of regular physical activity with PDFF_muscle.

Results from a linear regression model, adjusted for age, sex and BMI, outcome PDFF_muscle, exposure regular physical activity. ß‐coefficients denote a change in PDFF_muscle per unit (%).

PDFF: proton‐density fat‐fraction. Beta: ß‐coefficient. CI: confidence interval.

Table S3. Association of different risk factors with outcome IVDD.

Results from a logistic regression model, adjusted for age, sex and BMI, outcome IVDD. Continuous risk factors were standardized before modeling; ORs denote changes in risk for IVDD associated with an increase of one standard deviation of the continuous risk factor.

IVDD: intervertebral disc degeneration. CI: confidence interval. BMI: body mass index. PDFF: proton‐density fat‐fraction

Diallo T. D., Rospleszcz S., Fabian J., Walter S. S., Maurer E., Storz C., Roemer F., Rathmann W., Peters A., Jungmann P. M., Jung M., Bamberg F., and Kiefer L. S. (2023) Associations of myosteatosis with disc degeneration: A 3T magnetic resonance imaging study in individuals with impaired glycaemia, Journal of Cachexia, Sarcopenia and Muscle, 14, 1249–1258, 10.1002/jcsm.13192