Lumbar Disc Degeneration Is Linked to Dorsal Subcutaneous Fat Thickness at the L1–L2 Intervertebral Disc Level Measured by MRI

Reviewer 1 comments

Responses

Reviewer 1:

Thank you for the critical review of our manuscript, your suggestions have really improved the quality of our manuscript. We have now revised the paper taking into account all your comments.

Abstract:

Title: The title emphasizes a positive relationship between subcutaneous fat thickness and disc degeneration; however, the results of the study indicated a positive relationship only in the L1-2 intervertebral disc level and not in the other disc levels. Therefore, the title should be re-arranged, considering the study results or explaining the general concept of the study with methodology.

Thank you for pointing this out, We have now revised and re-arranged the tile to reflect the findings of the study as suggested. It says “Lumbar disc degeneration is linked to dorsal subcutaneous fat thickness at L1-L2 intervertebral disc level measured using magnetic resonance imaging (MRI)”

-Keywords: Additional keywords may be beneficial to find the article among the similar studies.

Thank you for this suggestion. We have now added additional keywords as suggested. It becomes “Keywords: degenerative disc disease; dorsal subcutaneous fat; magnetic resonance imaging; obesity; intervertebral disc“ see the title page, lines 28-29

Methods:

Comment 1:

Were there any patients with subcutaneous infection, abscess formation, or subcutaneous edema in the lumbar region? Were they excluded?

No, there were not. All cases were free from subcutaneous infection, abscess formation, or subcutaneous oedema in the lumbar region.  

Comment 2:

“The presence of DDD was assessed at each of 5 lumbar inter-vertebral levels, i.e. L1/2, L2/3, L3/4, L4/5, L5/S1, and the following items were utilized to confirm the presence of the disease- disc height lost, annular disc tears, disc bulge, spinal canal narrowing, disc desiccation, and disc extrusion” Instead of explaining the imaging findings used in the diagnosis, can you please give a more certain description of disc degeneration, and how these items were used in the diagnosis of a degenerative disc? Please use supportive data from the literature.

Thank you for your insightful observation. The Pfirrmann grading system, a well-established technique employed to evaluate the level of disc degeneration, was utilized for the evaluation of disc degeneration. This methodology takes into consideration various morphological and structural characteristics of the intervertebral discs, including disc height reduction, annular fissure, disc bulge, narrowing of the spinal canal, protrusion, and disc extrusion. In our study, participants lacking any of these aforementioned features were classified as the control group (without degenerative disc disease), whereas those displaying at least a single attribute constituted the case group (with degenerative disc disease). We have now considered this part in the method section to give a more certain description of disc degeneration, and how morphological and structural characteristics of the intervertebral discs were used in the diagnosis of a degenerative disc , lines 96-104

It reads “The Pfirrmann grading system, a well-established technique employed to evaluate the level of disc degeneration, was utilized for the evaluation of disc degeneration. This methodology takes into consideration various morphological and structural characteristics of the intervertebral discs, including disc height reduction, annular fissure, disc bulge, narrowing of the spinal canal, protrusion, and disc extrusion. Participants lacking any of these features were classified as the control group (without degenerative disc disease), whereas those displaying at least a single attribute constituted the case group (with degenerative disc disease).”    

Comment 3:

“Between December 2021 and August 2022, a total of 424 participants with MRI lumbar spine examinations were only recruited.” Did you mean all of the patients who had undergone MRIs in this period? Please indicate the methodology of the study clearly (retrospective or prospective? cross-sectional?).   

We apologize if this was unclear. Out of the 755 individuals who underwent MRI lumbar exams between December 2021 and August 2022, only 424 participants satisfied the inclusion criteria. We have now clarified this point in the method section (study population), line 89-91. It reads “A total of 424 participants, out of 755 individuals who received MRI lumbar spine examinations, were recruited between December 2021 and August 2022.”

To address reviewer comments on study design, we've now explicitly stated the study design in the Methods section (line 88). It reads “It is a single-center, retrospective cross-sectional study conducted at the MRI department of King Abdulaziz University Hospital, Jeddah, Saudi Arabia.”

Comment 3:

If this is your routine MR imaging protocol, please explain what other sequences are used in the routine lumbar MR imaging. I think the researchers have chosen T2-weighted sagittal-plane MR images for the measurements among other sequences. It will be important to discuss why sagittal-plane T2-weighted images were used instead of sagittal-plane T1-weighted images since T1-weighted images were commonly used to measure subcutaneous fat thickness in the literature. The benefits or disadvantages of using T2-weighted images to measure subcutaneous fat should be discussed in the discussion section.

We have now added more details about other MRI sequences used in the routine MRI lumbar spine protocol at our institution. See method section, lines 111-113

It reads “All images were acquired using the routine protocol:  Sagittal spin-echo T2 (TR 3900 ms, TE 100 ms), Sagittal Stair T2 (TR 3000 ms, TE 53 ms), Sagittal spin-echo T1 (TR 620 ms, TE 10 ms), Axial spin-echo T2 (TR 5000 ms, TE 94 ms). “

 T2-weighted sagittal-plane MR images were selected for the measurements of dorsal subcutaneous fat in the current study in accordance with prior studies that investigated the relationship between dorsal subcutaneous fat thickness and degenerative disc disease.[1,2] Currently, there is no standard approach for quantifying the thickness of the subcutaneous fat layer in the lumbar spine region. Additionally, it has been demonstrated that utilizing T1-weighted or T2-weighted MRI sequences yields closely associated outcomes, suggesting that either sequence can be employed to assess the thickness of subcutaneous fat without compromising the reliability of the estimated measurements.[3]

Comment 4:

“Using length functions of MRI console, dorsal subcutaneous fat thickness of L1 to S1 vertebral body levels was measured in mid-sagittal T2 weighted image where a line parallel to intervertebral disc space and extending from dorsal subcutaneous fat was drawn to measure the distance. Vertebral

body levels or inter-vertebral disc levels?

Figure 1 indicates the intervertebral disc levels as mentioned in the figure captions.

We apologize for this ambiguity. We meant intervertebral disc levels as indicated in the figure.1. We have now revised this point. It reads “Using length functions of MRI console, dorsal subcutaneous fat thickness of L1 to S1 intervertebral disc levels was measured in mid-sagittal spin-echo T2 weighted image where a line parallel to intervertebral disc space and extending from dorsal subcutaneous fat was drawn to measure the distance” see method section (Dorsal subcutaneous fat thickness measurements), lines 117-121.

Comment 5:

Why was a non-parametric test (Mann-Whitney U) used instead of a parametric test (An independent samples T-test?)? Was it about normality analysis? It would be better to explain this situation, and if there was no normal data distribution, please indicate the test used to analyze the data distribution (Shapiro-Wilk? Kolmogorov-Smirnov? etc.).

We apologize if this was unclear. The Shapiro-Wilk test and a histogram were utilized to judge the normality of numerical data which were found to be not normally distributed. Therefore, a non-parametric statistical test (Mann-Whitney U test) was implemented instead of an independent samples T-test to explore the differences between case and control groups. . We have now considered this in details as suggested. It reads “Statistical analysis was performed by JASP software (0.16.3). Data were presented as frequencies and proportions if they had categories. The Shapiro-Wilk test and a histogram were used to check the normality of numerical variables, which were found to be not normally distributed. Therefore, the continuous variables were described as medians and interquartile ranges. To investigate the differences between case and control groups in terms of population characteristics, Chi-squared test (X²) was used for categorical variables. Because the continuous data did not follow a normal distribution, we adopted a non-parametric statistical test namely Mann-Whitney U test. This test was used to investigate the differences between case and control groups in terms of population characteristics as well as the differences in dorsal subcutaneous fat thickness at all lumbar vertebral disc levels between males and females.” see method section (statistical analysis), lines 142-156.

Comment 6:

The method of the study indicates that 5 intervertebral levels were measured; however, six measurements are presented in Figure 1. Were T12-L1 levels of the patients included in the study as shown in the figure?

I apologize for this confusion. You are correct; only 5 intervertebral disc levels were measured, not 6 as indicated in Figure 1. This was an error, and the figure has now been revised accordingly. Please refer to the updated image in the method section on page 3, line 131 for clarification.

Comment 7:

“To ensure the reliability and accuracy of these measurements, the interclass correlation coefficient (ICC) of measurements for 30% of the study population was estimated for each vertebral body level”. Instead of 30% of the population, the whole study population would be better for finding more accurate ICC values.

We have now computed the interclass correlation coefficients for all measurements of dorsal subcutaneous fat thickness within the study population. Please refer to Table 1. In addition to that, the estimates of ICC within the text were revised accordingly. See results section, line 167

It reads “The repeatability of dorsal subcutaneous fat thickness measurements (ICC) between the two radiologists was excellent, at 0.99, 0.99, 0.98, 0.99, and 0.99 for L1-L2. L2-L3, L3-L4, L4-L5, and L5-S1 levels, respectively.”

Comment 8:

Figure:

L1-2 and L2-3 measurements are not parallel to the disc level in Figure 1. A re-arrangement in Figure 1 would better reflect the methodology indicated in the methods section.

Thank you for raising this out.  We’ve now adjusted the line so it's perfectly parallel to the intervertebral disc level, just as mentioned in the methodology. You can check out the updated image in the methods section on page 3, line 131, to see the change.

-Comment 9:

Language: A minor language editing is recommended.

We have made numerous revisions to several sentences within the manuscript in order to enhance the linguistic quality.

Results:

Comment 1:

Considering the use of the logistic regression analysis for the positive relationship in the L1-2 level [According to the logistic regression analysis, the odds for having DDD increased by 30% per one-millimeter increase of dorsal subcutaneous fat thickness at L1-L2 level (p = 0.03)] the association between subcutaneous fat thickness and disc degeneration was expressed in detail for this disc level. However, no relationship between subcutaneous fat thickness and disc degeneration was observed for other levels according to the results of the study. Please indicate this situation clearly in the results section.   

Thank you for pointing this out. We have now considered all insignificant relationships in the logistic regression analysis. (see table.4) Furthermore, the findings have been explicitly elucidated within the text. It says "According to the logistic regression analysis, there is a 30% increase in the likelihood of experiencing DDD for every one-millimeter increase in dorsal subcutaneous fat thickness at the L1-L2 level (p = 0.03). However, no significant association between subcutaneous fat thickness and disc degeneration was observed for the other levels (p ≥ 0.14)"; please refer to the result section, specifically lines 184-188. 

Discussion:

Comment 1:

As I mentioned before, it will be better to discuss why sagittal-plane T2-weighted images were used instead of sagittal-plane T1-weighted images. The benefits or disadvantages of using T2-weighted images to measure subcutaneous fat should be discussed in the discussion section.

T2-weighted sagittal-plane MR images were selected for the measurements of dorsal subcutaneous fat in the current study in accordance with prior studies that investigated the relationship between dorsal subcutaneous fat thickness and degenerative disc disease.[1,2] Currently, there is no standard approach for quantifying the thickness of the subcutaneous fat layer in the lumbar spine region. Additionally, it has been demonstrated that utilizing T1-weighted or T2-weighted MRI sequences yields closely associated outcomes, suggesting that either sequence can be employed to assess the thickness of subcutaneous fat without compromising the reliability of the estimated measurements.[3]

Comment 2:

The number of patients equal to or older than 40 years was statistically much more than the younger ones. How will this situation affect the results? Considering this situation, the effect of age should be discussed in more detail.

Indeed, we agree that our findings on the significantly stronger relationship between DSFT and DDD among the younger age group may have been influenced by a greater number of participants who were older. As subcutaneous fat has been shown to decrease with age, older individuals are more likely to have modest amounts of dorsal subcutaneous fat, which may bias the substantial function of fat toward younger age groups. Therefore, a more balanced sample size is needed to confirm this relationship.  We have now discussed this part in the discussion section. It reads “Research revealed that subcutaneous fat thickness decreases with aging.[4] This decline with age means that older participants tend to have less dorsal subcutaneous fat thickness, which potentially skewing its importance in function towards younger populations in the current study. Given this age-bias, our findings linking DDD to dorsal subcutaneous fat thickness in younger individuals should be treated with caution. A more balanced age distribution in future studies is crucial to confirm this relationship”

Reviewer 2 comments

Responses

Reviewer 2:

Thank you for the valuable time and effort you have put into reviewing our manuscript. We now revised the paper taking into account all your comments.

Abstract:

Comment1:

The “DSFT” appeared without a full-term introduction.

We appreciate you bringing this to our attention. We have addressed your suggestion and the abstract now states "Therefore, this study aims to investigate the association between dorsal subcutaneous fat thickness (DSFT) of the lumbar spine, an alternative body fat distribution metric, and DDD." (line 15)

Comment2:

It is stated that DSFT was measured in the mid-sagittal T2WI, but information about the specific level or range from which to which level it was measured is missing. This is considered crucial information, and I recommend adding it.

We have now addressed this comment by adding more details about the specific level or range from which-to-which level DSFT was measured in the abstract section. It reads “Using length functions of MRI console, DSFT of L1 to S1 intervertebral disc levels was measured in mid-sagittal spin-echo T2 weighted image.” see abstract, lines 17-19

Comment 3:

Please provide information in the abstract about the statistical methods conducted and their results before the use of logistic regression

We have now expanded the abstract to include details of statistical tests employed, using the Mann-Whitney U and Chi-squared tests to assess case-control group differences. It reads “The Mann-Whitney U test and Chi-squared test (X2) were utilized to examine any variations between the case and control groups.” See abstract section, lines 22-23

Introduction:

Comment 1:

 Page 2 line 51~: several previous studies are introduced that elucidate the association between BMI or MRI adiposity measurements and DDD. Among these, the paper titled "Fat quantification of multifidus muscle using T2-weighted Dixon: which measurement methods are best suited for revealing the relationship between fat infiltration and herniated nucleus pulposus" in Skeletal Radiol 49, 263–271 (2020) appears to be particularly relevant to this topic. Referring to this paper in the introduction could provide readers with additional information.

Thank you for highlighting this point. We have now considered the suggested article in the introduction section. Lines 59-61.

It reads “A study conducted by Lee et al. suggests a potential link between herniated nucleus pulposus and specific fat measurements in the spine. The study involved 108 patients who underwent MRI scans of their spines due to lower back pain. Their findings revealed a significant association between higher coronal 2D fat fraction and the presence of herniated nucleus pulposus (HNP).[1]” See introduction section, lines 57-61

Methods:

Comment 1:

2.1 Study population : the definition of DDD is provided, but many non-consensus findings are noticeable (e.g., disc desiccation), and there is no mention of disc protrusion. Please consider revising the definition by referring to "Lumbar disc nomenclature: version 2.0: Recommendations of the combined task forces of the North American Spine Society, the American Society of Spine Radiology, and the American Society of Neuroradiology. Spine J. 2014;14(11):2525-2545," to incorporate commonly accepted findings.

We are grateful for bringing this to our attention. We have now incorporated the provided source into the revised method section (lines 96-107)

It reads “The Pfirrmann grading system, a well-established technique employed to evaluate the level of disc degeneration, was utilized for the evaluation of disc degeneration. This methodology takes into consideration various morphological and structural characteristics of the intervertebral discs, including disc height reduction, annular fissure, disc bulge, narrowing of the spinal canal, protrusion, and disc extrusion. Participants lacking any of these features were classified as the control group (without degenerative disc disease), whereas those displaying at least a single attribute constituted the case group (with degenerative disc disease).”      

Comment 2:

 Lumbar MRI: The introduced MRI information on Lumbar MRI is insufficient. Please add details regarding the MRI equipment manufacturer used in your hospital, sequence information, plane details, and other relevant information pertaining to lumbar spine MRI conducted in your institution.

We have now considered more details about MRI equipment, and sequences used at out institution as suggested. It reads “All patients have undergone a lumbosacral spine MRI scan on a high-field strength 3.0 tesla machine (Magentom, Siemens, USA). The scan was performed while the patients were in a supine position using a multichannel phased array spine surface coil for optimal imaging. All images were acquired using the routine protocol:  Sagittal spin-echo T2 (TR 3900 ms, TE 100 ms), Sagittal Stair T2 (TR 3000 ms, TE 53 ms), Sagittal spin-echo T1 (TR 620 ms, TE 10 ms), Axial spin-echo T2 (TR 5000 ms, TE 94 ms)..” See method section, lines 108-113

Comment 3:

Dorsal subcutaneous fat thickness measurement: it is mentioned that ICC was conducted on only 30% of the data for dorsal subcutaneous fat thickness measurement. However, it is generally understood that ICC should be performed for all cases. If possible, please measure ICC using the entire dataset.

We have now computed the interclass correlation coefficients for all measurements of dorsal subcutaneous fat thickness within the study population. Please refer to Table 1. In addition to that, the estimates of ICC within the text were revised accordingly. See results section, line 167

It reads “The repeatability of dorsal subcutaneous fat thickness measurements (ICC) between the two radiologists was excellent, at 0.99, 0.99, 0.98, 0.99, and 0.99 for L1-L2. L2-L3, L3-L4, L4-L5, and L5-S1 levels, respectively.”

Comment 4:

Statistical analysis: There is no mention of ICC in the statistical analysis section. Additionally, please provide interpretations for ICC values. If needed, refer to Skeletal Radiol 49, 263–271 (2020) for guidance.

Thank you for highlighting this point. We have now mentioned the ICC in the statistical analysis section with the interpretation of ICC values. It reads “To assess how consistently different two observers measured the thickness of dorsal subcutaneous fat for each intervertebral disc level, we calculated inter-observer reliability using interclass correlation coefficients (ICCs) for all study population (see table.1). A two-way random model with absolute agreement was used. Scores were interpreted as: poor (< 0.40), fair (0.40-0.59), good (0.60-0.74), or excellent (0.75-1.00).” see method section, lines 159-164

Comment 5:

Statistical analysis: is there a specific reason for conducting the Mann-Whitney U test instead of the Student t-test, especially when the sample size is sufficient?

We apologize if this was unclear. The Shapiro-Wilk test and a histogram were utilized to judge the normality of numerical data which were found to be not normally distributed. Therefore, a non-parametric statistical test (Mann-Whitney U test) was implemented instead of an independent samples T-test to explore the differences between case and control groups. . We have now considered this in details suggested. It says “Statistical analysis was performed by JASP software (0.16.3). Data were presented as frequencies and proportions if they had categories. The Shapiro-Wilk test and a histogram were used to check the normality of numerical variables, which were found to be not normally distributed. Therefore, the continuous variables were described as medians and interquartile ranges. To investigate the differences between case and control groups in terms of population characteristics, Chi-squared test (X²) was used for categorical variables. Because the continuous data did not follow a normal distribution, we adopted a non-parametric statistical test namely Mann-Whitney U test. This test was used to investigate the differences between case and control groups in terms of population characteristics as well as the differences in dorsal subcutaneous fat thickness at all lumbar vertebral disc levels between males and females.” see method section (statistical analysis), lines 140-154.

Comment 6:

[Figure]

I think it would be good if you could show an representative case that supports your hypothesis.

Thank you for highlighting this point, we have now considered a representative case that supports our hypothesis. See figure.1 (a), page 3.

Results:

Comment 1:

Table 3: Similar to the question raised in the statistical analysis section, could you please clarify the reason for conducting statistical analysis on the difference between DSFT and gender? It raises concerns about whether this analysis might deviate from the main topic. If necessary, it would be advisable to separate the layout of tables 3 and 4,5.

The current study highlights a potential association between dorsal subcutaneous fat thickness and degenerative disc disease specifically in women. This finding warrants further investigations to determine the differences in the dorsal subcutaneous fat thickness between males and females. If the fat thickness is greater in women as reported in the study, this might provide a more valuable indicator of obesity-related lumbar spine issues in women compared to men. Such insights could ultimately lead to more tailored and effective strategies for preventing and treating spinal problems in women.

We have now clarified this point. It states” The differences on dorsal subcutaneous fat thickness by gender was investigated to explore the hypothesis of women being more effective in storing fat subcutaneously compared to men. Doing so, it would help to reveal whether this fat layer provides a more helpful measure of obesity-related lumbar spine problems in females compared to males.”  

Comment 2:

Table 4,5: Please separate the cells for odds ratio and p-value in these tables

We have now separated cells to represent odds ratio and p-values individually in the tables 4 and 5. See tables 4 and 5.

Discussion:

Comment 1:

Page 6, line 182~ The fact that the subcutaneous fat layer thickness at L1-2 is associated with other diseases is very interesting. Please provide a more detailed information by elaborating on previous studies that mention the significance of L1-2 fat thickness, even in conditions other than DDD.

Thank you for raising this point. We have now elaborated the previous studies that reported the significant role of DSFT at L1-L2 level in the prognostication of several spinal disorders. See discussion, lines 228-233

It reads” Prior studies supported the potential clinical utility of the dorsal subcutaneous fat thickness at L1-L2 level in the prognostication of several spinal disorders.[2,3] A recent study conducted by Ozcan-Eksi et al. reported that dorsal subcutaneous fat thickness measured from MRI and CT at LI-L2 level is significantly associated with DDD. [3]West et al. study found that dorsal subcutaneous fat thickness at L1-L2 level can predict oedema in the lumbosacral subcutaneous fat tissue.[2]”

Comment 2:

While I agree that DSFT serves as an objective tool for measuring obesity, there is uncertainty regarding the cost-effectiveness of conducting spine MRI for body fat measurement. I am just curious about whether It is reasonable within the medical insurance of your institution or country.  

Although MRI offers excellent soft tissue differentiation and can measure subcutaneous fat thickness, its high cost compared to alternatives like ultrasound raises concerns about its applicability for this purpose. Typically, medical insurance covers treatments and procedures deemed medically necessary based on evidence-based guidelines. As spine MRI for measuring dorsal subcutaneous fat thickness is not a standard practice for obesity assessment in Saudi Arabia, and MRI is usually reserved for specific medical indications due to its cost and resource constraints, employing it solely for fat thickness measurement, particularly in an obesity context, is likely deemed an unnecessarily expensive approach. We have now considered this point in the discussion section, lines 286-294, it reads “Although MRI offers excellent soft tissue differentiation and can measure subcutaneous fat thickness, its high cost compared to alternatives like ultrasound raises concerns about its applicability for this purpose. Typically, medical insurance covers treatments and procedures deemed medically necessary based on evidence-based guidelines. As spine MRI for measuring dorsal subcutaneous fat thickness is not a standard practice for obesity assessment in Saudi Arabia, and MRI is usually reserved for specific medical indications due to its cost and resource constraints, employing it solely for fat thickness measurement, particularly in an obesity context, is likely deemed an unnecessarily expensive approach.”

Reviewer 2 comments

Responses

Reviewer 2:

Thank you for the valuable time and effort you have put into reviewing our manuscript. We now revised the paper taking into account all your comments.

Abstract:

Comment1:

I think the title could just say 'MRI' as an abbreviation

Thank you for highlighting this point. we've now replaced the full term "magnetic resonance imaging" with the widely recognized abbreviation "MRI" in the title. It reads “Lumbar disc degeneration is linked to dorsal subcutaneous fat thickness at the L1-L2 intervertebral disc level measured by MRI”

Comment2:

How about applying 'DSFT' to the whole text as written in the abstract?

Thank you for raising this point. We have now considered the abbreviation “DSFT” throughout the manuscript as suggested. Written in blue colour.

Comment 3:

Please recheck the entire text for typos and spacing

We have now rechecked and fixed the entire text for typos and spacing as suggested.

Introduction:

Comment 1:

Please confirmed that Lee's study introduced in the Introduction studied the fat fraction of 'paraspinal muscles'.

Thank you for highlighting this point. We have now clarified that Lee and colleagues investigated the relationship of the fat fraction of 'paraspinal muscles with herniated nucleus pulposus. Lines 59-61.

It now reads “A study conducted by Lee et al. suggests a potential link between herniated nucleus pulposus and the fat content of paraspinal muscles. The study involved 108 patients who underwent MRI scans of their spines due to lower back pain. Their findings revealed a significant association between a higher coronal 2D fat fraction of paraspinal muscles and the presence of herniated nucleus pulposus (HNP).[13]”