Dyslipidemia Is Positively Associated with Chronic Low Back Pain in Korean Women: Korean National Health and Nutrition Examination Survey 2010–2012

Round 1

Reviewer 1 Report

This study aimed to investigate the association between certain chronic diseases and low back pain (LBP) among Koreans. I have the following comments:

Abstract

Line 15: Since the definition of LBP was confined only to those who experienced LBP for at least 3 months, the authors may consider changing the outcome to chronic LBP because people with acute and subacute LBP were not included. Similar changes should be made to the title and across the manuscript.  

Line 20: Associated instead of correlated.

Lines 21-22: Given the cross-sectional design of this study, it’s presumptive to claim a causal relationship. Thus, claiming that dyslipidemia treatment can reduce LBP risk is speculative.

Introduction

General comment: English editing is required.

Lines 30-31: This study aims to investigate Koreans. The authors should report the financial burden of LBP among Koreans not in Sweden. Further, the Swedish findings are outdated.

Lines 45-47: The prevalence of aortic aneurysm globally is lower than 1% making this explanation very unlikely. Further, Ref 5 that was used by the authors is a non-peer reviewed online page that did not even mention this explanation. On the other hand, a previous Korean study using KNHANES showed a negative association between hypertension and LBP (PMID: 26393797).

Line 59: Correlation is a statistical approach. The authors should not use it instead of association.

General comment: The authors did not clarify why the analysis was restricted to women. Why the authors did not include men and women, investigate sex interaction, and stratify the results by sex if needed.

Methods

Line 73: The authors should have excluded pregnant women and those who have given birth during the past few months.

Figure 1: Given the very large number of excluded participants for having missing values (>45%), the authors should compare the baseline characteristics between the included and the excluded in a supplementary analysis. In the case there are significant differences between both groups, this should be admitted in the limitation section because it undermines the representativeness of the study.

Lines 119-135:

A: There are several inconsistencies in the definitions of the exposures. First, the authors claimed to stratify blood pressure into 3 groups, yet they described 2 groups only (normal [<120/80 mmHg] and hypertension [=or>140/90]). Further, the analysis included 2 groups only without mentioning the high normal blood pressure group; however, the total number of participants was correct; suggesting that high normal and optimal blood pressure groups were merged. Merging the high normal and optimal blood pressure groups into one group may probably attenuate the association and lead to a differential misclassification bias.

B: Similarly, according to the authors, diabetes was stratified into 2 groups only, yet the normal group included those with FBG < 100mg/dL while the diabetes group included those with FBG =>126 mg/dL.

C: Furthermore, the authors based their definitions for hypertension, diabetes, and dyslipidemia on both lab findings and self-report. How did the authors handle subjects with inconsistencies, normal labs but positive report or abnormal labs and negative report.

Results

Table 2

BMI groups should be rearranged: <18.5, 18.5-24.9 (Reference), and =>25 kg/m2. Underweight is associated with sarcopenia which is a risk factor for LBP; so, it should not be a reference group.

Table 3

Multivariable not multivariate regression. They are different.

The variables included in the regression models should be shown in the footnotes.

Given the huge differences between the results of the univariate and the multivariable regression analyses, more than one model should be included. For example, model I adjusted for age, model II further adjusted for lifestyle, and model III further adjusted for other clinical conditions including obesity.

General comment: Since dyslipidemia is strictly related to obesity and obesity is the most significant risk factor for LBP, it is probable that dyslipidemia led to LBP via obesity and this association is not independent. Although the authors might have adjusted their results for obesity, this adjustment is not enough. Stratifying the results by the obesity status is very important.

Discussion

Lines 211-214: The authors cannot claim these associations because they were only indicated using univariate analyses.

Lines 215-247: They are irrelevant to the current study. The aim of the study, as reported by the authors, is to examine the association between chronic diseases and LBP. These lines should be deleted. In addition, these findings were not confirmed by multivariable-adjusted regression analyses.

Line 250: Unlike the claim of the authors, Ref 10 showed an association between hypertension and severe LBP only but not the overall LBP, while Ref 11 showed a positive association in men but not women (the target population in this study). The absence of a positive association in this study may be attributed to inadequate categorization.

Lines 346-347: Did the authors exclude all participants with any missing values or only participants with missing values in the exposure and the outcome?

General comment: Another major limitation is that LBP was diagnosed based on a simple yes/no question. No data about severity, duration, or medications were collected. Thus, a dose-response association could not be detected.

Lines 357-359: Speculative conclusion.  

Moderate editing.

Author Response

Thank you so much for the great feedback. I will post a description of the modifications you mentioned in the attachment. 

Author Response File: Author Response.pdf

Reviewer 2 Report

The study by Kim and Lee explores the association between low back pain (LBP) and chronic diseases such as hypertension, diabetes, and dyslipidemia, using data from the Korean National Health and Nutrition Examination Survey (KNHANES) from 2010 to 2012. The focus is on a specific demographic, Korean women, who reported LBP for more than three months.

Methodology: The study's methodology is robust, employing two-stage stratified cluster sampling for data extraction and using multivariate logistic regression analysis. The analysis is comprehensive, adjusting for various confounding factors including age, weight, height, waist circumference, individual income, and lifestyle factors such as exercise and diet.

Key Findings: Among 5961 participants, dyslipidemia showed a significant positive association with LBP. This association persisted even after adjusting for confounding variables. However, no correlation was found between LBP and other chronic diseases like hypertension and diabetes.

Suggestions for improving the paper:

Detailed Analysis of Confounding Factors: Strengthen the analysis by providing more detailed information on how obesity and related factors (like BMI and waist circumference) were controlled in the statistical model. This would clarify the independent effect of dyslipidemia on LBP.

Mechanistic Explanations: Include a discussion section that hypothesizes potential physiological or biochemical mechanisms linking dyslipidemia to LBP, based on existing literature. This could provide a theoretical basis for the observed association.

Subgroup Analysis: Conduct subgroup analyses to see if the association between dyslipidemia and LBP varies across different age groups, BMI categories, or other relevant factors within the study population.

Address Statistical Limitations: More thoroughly discuss the limitations of the statistical methods used, such as the potential impact of the cross-sectional design on the results and how this might affect the interpretation of the data.

Enhance Discussion on Generalizability: Expand the discussion on the generalizability of the findings, particularly in relation to how these results might apply to other populations beyond Korean women.

Comments for author File: Comments.pdf

Author Response

Thank you so much for the great feedback. I will post a description of the modifications you mentioned in the attachment. 

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript is substantially improved. The authors did their best to address reviewers' comments; therefore, it could proceed to publication. The data analysis and results sections provide significant insights into the mechanisms linking dyslipedemia and obesity with chronic low back pain. Limitations are sufficiently discussed. The only research design issue of the study is its cross-sectional nature. 

The term "multivariate" is the correct one (instead of multivariable), therefore please amend accordingly throughout the manuscript.

Author Response

I have responded to your valuable comments via word files, thank you. 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thank you for addressing my comments.

Please use the term multivariable regression analysis instead of multivariate regression analysis. 

Hidalgo B, Goodman M. Multivariate or multivariable regression? Am J Public Health. 2013 Jan;103(1):39-40.

Minor editing is needed.

Author Response

Thank you for the opportunity to make a minor revision. 

Please refer to the attachment for answers to your good advice. 

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors fixed the paper following my instructions

Author Response

Thank you for the opportunity to make a minor revision. 

Please refer to the attachment for answers to your good advice. 

Author Response File: Author Response.pdf