Is LDL cholesterol associated with long-term mortality among primary prevention adults? A retrospective cohort study from a large healthcare system

Abstract

Key Aspects

• Objective: The primary objective of this study is to evaluate the relationship between low-density lipoprotein cholesterol (LDL-C) and long-term mortality in a real-world population of primary prevention-type adults not on lipid-lowering therapy. This addresses the existing conflicting results in the literature regarding this association.
• Study Design: The study utilizes a retrospective cohort design, analyzing electronic medical record data from the University of Pittsburgh Medical Center healthcare system. The data spans from January 4, 2000, to December 31, 2022.
• Study Population: The study population consists of adults aged 50-89 years without diabetes who were not on statin therapy at baseline or within one year. Participants were classified as primary prevention-type patients. To mitigate potential reverse causation, patients who died within one year or had baseline total cholesterol (T-C) ≤120 mg/dL or LDL-C <30 mg/dL were excluded.
• Exposure: The main exposure measure is baseline LDL-C, categorized into six groups: 30–79, 80–99, 100–129, 130–159, 160–189, and ≥190 mg/dL.
• Outcome: The primary outcome measure is all-cause mortality, with follow-up starting 365 days after the baseline cholesterol measurement.
• Key Findings: The study found a U-shaped relationship between LDL-C categories and mortality. The lowest risk for long-term mortality was observed in the LDL-C range of 100–189 mg/dL. Adjusted mortality hazard ratios (HRs) compared to the referent group (LDL-C 80–99 mg/dL) showed significantly lower risk for the 100-129, 130-159, and 160-189 mg/dL groups. In contrast, T-C/HDL cholesterol and triglycerides/HDL cholesterol ratios were independently associated with long-term mortality.
• Conclusions: The study concludes that among primary prevention-type patients aged 50–89 years without diabetes and not on statin therapy, the lowest risk for long-term mortality exists in the wide LDL-C range of 100–189 mg/dL, which is much higher than current recommendations. The authors suggest minimal consideration should be given to LDL-C concentration when counseling these patients.

Strengths

• Large Sample Size and Long-Term Follow-Up

  The study includes a large sample of 177,860 patients with a mean follow-up of 6.1 years, providing robust data for analysis.

  "177 860 patients with a mean (SD) age of 61.1 (8.8) years and mean (SD) LDL- C of 119 (31) mg/dL were evaluated over a mean of 6.1 years of follow- up." (Page 1)
• Mitigation of Reverse Causation

  The study design addresses potential reverse causation by excluding patients who died within one year of baseline cholesterol measurement or had exceptionally low T-C or LDL-C levels.

  "To mitigate potential reverse causation, patients who died within 1 year or had baseline total cholesterol (T- C) ≤120 mg/dL or LDL- C <30 mg/dL were excluded." (Page 1)
• Real-World Evidence

  The study utilizes electronic medical record data, reflecting a real-world clinical setting and enhancing the generalizability of the findings.

  "We evaluated this relationship in a real- world evidence population of adults." (Page 1)

Suggestions for Improvement

• Clarify Cause-Specific Mortality

  This medium-impact improvement would enhance the study's clinical relevance and impact on patient care guidelines. The Abstract section is the appropriate place for this clarification as it provides the first impression of the study's scope and limitations.

  "The analysis was limited to all- cause mortality and thus was unable to assess cause- specific mortality." (Page 1)

  Implementation: Add a sentence explicitly stating that the study was unable to assess cause-specific mortality due to data limitations, and briefly mention the implications of this limitation. For example: "Due to the nature of the data, cause-specific mortality could not be assessed, limiting our ability to determine the impact of LDL-C on specific causes of death."

• Elaborate on Clinical Implications

  This high-impact improvement would strengthen the abstract's impact by directly addressing the clinical implications of the findings. The Abstract is the ideal location for this because it is often the only part of the paper that many readers will see, making it crucial for conveying the study's practical significance.

  "For counselling these patients, minimal consideration should be given to LDL- C concentration." (Page 1)

  Implementation: Expand the conclusion to include a brief discussion on how the findings might influence clinical practice, particularly regarding patient counseling and treatment decisions. For example: "These findings suggest that current guidelines recommending aggressive LDL-C lowering in primary prevention may need reevaluation. Clinicians should consider a more holistic approach, focusing on overall cardiovascular risk factors rather than solely on LDL-C levels."

• Specify Statistical Adjustments

  This low-impact improvement would enhance the abstract's transparency and allow for a more nuanced understanding of the presented results. While the full methods are detailed in the paper, briefly mentioning the key covariates adjusted for in the analysis would improve the abstract's stand-alone interpretability.

  "Adjusted mortality HRs as compared with the referent group of LDL- C 80–99 mg/dL were: 30–79 mg/ dL (HR 1.23, 95% CI 1.17 to 1.30)," (Page 1)

  Implementation: Add a phrase to the results section indicating the key factors adjusted for in the analysis. For example: "Adjusted mortality HRs, controlling for age, sex, race, and other cardiovascular risk factors, as compared with the referent group..."

Introduction

Key Aspects

• Context and Motivation: The Introduction establishes the context of the study by highlighting the prevailing belief that high total cholesterol (T-C), particularly low-density lipoprotein cholesterol (LDL-C), is a primary cause of atherosclerotic cardiovascular disease (ASCVD), a major component of heart disease (HD). It notes that HD is the leading cause of death in the USA and that there is a near-universal belief that "lower is better" when it comes to LDL-C levels, with an optimal level often cited as at or below 100 mg/dL.
• Clinical Guidelines and Practice: The section discusses current clinical guidelines, specifically those from the American College of Cardiology (ACC), which implicate elevated LDL-C as a cause of ASCVD and advocate for the use of statins in primary prevention for patients with an estimated 10-year risk of ASCVD ≥7.5%. The ACC-ASCVD Risk Estimator, a tool used to estimate this risk, is mentioned, along with the observation that it may classify many older males as being at intermediate or high risk even with 'normal' risk factors.
• Conflicting Evidence and Research Gap: The authors present conflicting evidence that challenges the "lower is better" LDL-C paradigm. They cite meta-analyses suggesting that high LDL-C is associated with only a small increased risk of ASCVD or premature mortality and that the absolute risk reductions from lipid-lowering therapies are modest. They also point to the occurrence of acute coronary syndromes in patients with 'normal' LDL-C and the paradox that women, who generally have higher LDL-C levels, tend to have lower coronary HD mortality than men.
• Life Insurance Medicine Perspective: The Introduction incorporates insights from life insurance medicine, which focuses on predicting mortality hazards. It highlights that the T-C/HDL-C ratio is considered a better predictor of all-cause mortality than LDL-C in this field. Selected life insurance underwriting guidelines are presented to illustrate that high T-C levels may still qualify for favorable policies if the T-C/HDL-C ratio is within an acceptable range.
• Study Objective and Rationale: The authors state their objective: to critically analyze the relationship between LDL-C and long-term mortality risk in adults within a large, 'real-world' healthcare system. They aim to evaluate this association among primary prevention-type adults without diabetes aged 50-89 years, without a specific focus on statin therapy. This objective is motivated by the conflicting evidence and the need for a more nuanced understanding of the LDL-C-mortality relationship.

Strengths

• Clear Problem Statement

  The Introduction effectively establishes the research problem by highlighting the conflict between prevailing beliefs about LDL-C and emerging evidence.

  "Among primary prevention- type adults not on lipid- lowering therapy, conflicting results exist on the relationship between low- density lipoprotein cholesterol (LDL- C) and long- term mortality." (Page 1)
• Comprehensive Background

  The section provides a thorough overview of relevant background information, including clinical guidelines, meta-analysis findings, and insights from life insurance medicine.

  "A near- universal but not absolute belief3 is that high total choles- terol (T- C), low- density lipoprotein choles- terol (LDL- C) in particular (the so- called ‘bad’ cholesterol), is a root cause of ASCVD" (Page 1)
• Well-Defined Research Gap

  The Introduction clearly articulates the research gap by pointing out the need for a critical analysis of the LDL-C-mortality relationship in a real-world setting.

  "call for a critical appraisal and analysis of the relationship between LDL- C and long- term risk of mortality in adults." (Page 2)

Suggestions for Improvement

• Strengthen Rationale for Primary Prevention Focus

  This medium-impact improvement would enhance the Introduction's justification for focusing on primary prevention. While the rationale is touched upon, explicitly connecting it to the broader debate surrounding LDL-C targets and statin use in primary prevention would further strengthen the study's relevance and potential impact.

  "The analysis did not focus on the use of statin therapy for primary prevention." (Page 2)

  Implementation: Add a paragraph that elaborates on the specific controversies and uncertainties surrounding LDL-C lowering in primary prevention, emphasizing the need for more evidence in this specific population. For example: "The debate surrounding optimal LDL-C targets is particularly relevant in primary prevention, where the benefits of aggressive lowering are less clear-cut compared to secondary prevention. Current guidelines vary, and the decision to initiate statin therapy often involves complex risk-benefit considerations. This study aims to contribute to this ongoing debate by providing real-world evidence on the association between LDL-C and mortality in a large cohort of primary prevention individuals."

• Elaborate on Limitations of Existing Studies

  This low-impact improvement would provide a more nuanced understanding of the existing literature. While the Introduction mentions conflicting evidence, briefly discussing the limitations of previous studies (e.g., sample size, study design, population characteristics) would further justify the need for the current study and highlight its unique contribution.

  "meta-analyses indicate that high LDL-C is associated with at most a small increased absolute risk of ASCVD or premature mortality." (Page 2)

  Implementation: Include a brief discussion of the limitations of the cited meta-analyses and observational studies. For example: "While these meta-analyses provide valuable insights, they are often limited by the heterogeneity of included studies, varying definitions of LDL-C categories, and potential residual confounding. Furthermore, many observational studies may not fully account for factors such as socioeconomic status and access to healthcare, which can influence both LDL-C levels and mortality risk. The present study aims to address some of these limitations by analyzing a large, well-defined cohort within a single healthcare system."

• Clarify the Novelty of the Study

  This medium-impact improvement would further highlight the study's contribution to the field. While the Introduction implies the study's novelty, explicitly stating what makes this study unique compared to previous research would further strengthen its impact and rationale.

  "Therefore, within a large, ‘real-world’ healthcare system, we evaluated the association between LDL-C and all-cause long-term mortality among primary prevention-type adults" (Page 2)

  Implementation: Add a sentence or two that explicitly state the novel aspects of the study. For example: "Unlike previous studies that have primarily relied on data from clinical trials or diverse populations, this study leverages a large, real-world dataset from a single healthcare system, allowing for a more detailed examination of the LDL-C-mortality relationship in a primary prevention setting. Additionally, the study's long-term follow-up and comprehensive data collection enable a robust assessment of potential confounders and effect modifiers."

Non-Text Elements

Supplement Table 1. ASCVD 10-Year Risk Calculations for Primary Prevention* by...

Full Caption

Supplement Table 1. ASCVD 10-Year Risk Calculations for Primary Prevention* by Age, Race, and Sex

Figure/Table Image (Page 15)

First Reference in Text

Risk of ASCVD is often estimated using the online ACC-ASCVD Risk Estimator, and as seen in online supplemental table 1, all males ages 59 and older even in the presence of 'normal' ASCVD risk factors (lipids included) may be classified at intermediate or high risk of ASCVD, and thus candidates for LDL-C lowering therapy.

Description

• Table purpose: This table shows the estimated 10-year risk of developing ASCVD for different groups of people, categorized by age, race, and sex. ASCVD stands for atherosclerotic cardiovascular disease, which includes conditions like heart attacks and strokes caused by plaque buildup in the arteries. 'Primary prevention' means that these calculations are for people who have not yet had an ASCVD event. It's like trying to predict the chances of someone getting these diseases in the next 10 years, based on their current characteristics.
• Table structure: The table is organized with rows representing different age groups and columns representing different combinations of race and sex (e.g., White Male, Black (AA) Male, White Female, Black (AA) Female). Each cell in the table likely contains a percentage value, which represents the estimated 10-year risk of ASCVD for that particular group. For example, one cell might show the risk for white males aged 50-59, while another might show the risk for black females aged 60-69.
• Risk categories: The table also categorizes the estimated risk into different levels, such as 'Low,' 'Borderline,' 'Intermediate,' and 'High.' These categories provide a way to interpret the percentage values and understand the relative risk for each group. For example, a risk of less than 5% might be considered 'Low,' while a risk greater than 20% might be considered 'High.'
• Risk factors used: The footnote to the table specifies that the risk calculations are based on certain assumptions about the individuals' characteristics, including total cholesterol, LDL cholesterol, HDL cholesterol, systolic blood pressure, diastolic blood pressure, smoking history, and medication use. These are all factors known to influence the risk of ASCVD. It's like saying, 'these risk estimates are based on people with these specific characteristics, which are considered 'normal' or average values.'
• Source of calculations: The reference text mentions that the risk estimates are based on the ACC-ASCVD Risk Estimator, which is a widely used online tool developed by the American College of Cardiology and the American Heart Association. This tool uses a specific formula, or algorithm, to calculate the 10-year risk of ASCVD based on an individual's characteristics.

Scientific Validity

• Validity of the ACC-ASCVD Risk Estimator: The ACC-ASCVD Risk Estimator is a well-established tool based on data from large cohort studies. However, its accuracy may vary depending on the population being assessed. The tool has been validated primarily in US populations, and its applicability to other populations may be limited.
• Appropriateness of 'normal' values: The use of 'normal' values for risk factors in the table footnote is a simplification. While these values may represent average or guideline-recommended levels, they do not necessarily reflect the actual distribution of risk factors in the population. The specific values used should be clearly defined and justified.
• Limitations of risk prediction: It is important to recognize that risk prediction models are not perfect and have limitations. They provide estimates based on population averages and may not accurately predict individual risk. Other factors not included in the model may also influence an individual's risk of ASCVD.
• Relevance to the study's research question: The table provides context for the study by illustrating how the ACC-ASCVD Risk Estimator categorizes individuals based on their risk factors. However, the direct relevance to the study's main research question about the relationship between LDL-C and mortality may be limited, as the table does not specifically address this relationship.

Communication

• Clarity of caption: The caption clearly states the content and purpose of the table, including the outcome being assessed (ASCVD 10-year risk), the target population (primary prevention), and the variables used for categorization (age, race, and sex).
• Organization and layout: The table is well-organized, with clear headings for each column and row. The use of separate columns for different race and sex groups facilitates comparison. The specific layout and formatting details are not provided in the text but should be designed for optimal readability.
• Use of abbreviations: The table uses abbreviations (e.g., ASCVD, AA). These abbreviations are defined in either the caption or the footnote.
• Accessibility to non-experts: The table presents relatively straightforward data that can be understood by a broad audience. However, a brief explanation of the concept of ASCVD risk and the purpose of risk estimation in the introduction or discussion section could enhance comprehension for non-experts.
• Completeness of information: The footnote provides important information about the assumptions used in the risk calculations. However, it could be further improved by specifying the exact values used for each risk factor (e.g., total cholesterol = 190 mg/dL). The source of the risk estimation formula (ACC-ASCVD Risk Estimator) should also be explicitly stated in the table or the accompanying text.

Supplement Table 2. Maximum/Range of Total Cholesterol (T-C) Values Along with...

Full Caption

Supplement Table 2. Maximum/Range of Total Cholesterol (T-C) Values Along with T-C to HDL-C Cholesterol Ratios for Different Life Insurance Underwriting Categories

Figure/Table Image (Page 16)

First Reference in Text

As seen in online supplemental table 2, T-C and HDL-C are used jointly in policy underwriting, whereas LDL-C is not used, and lipid-lowering therapy is not emphasised.

Description

• Table purpose: This table shows the acceptable ranges of total cholesterol and the ratio of total cholesterol to HDL cholesterol for different categories used in life insurance underwriting. Life insurance underwriting is the process by which insurance companies assess the risk of insuring an individual, which then determines the premium (the cost of the insurance). In this case, the table is showing how cholesterol levels are used to help determine that risk.
• Cholesterol measures: The table focuses on two main cholesterol measures: Total Cholesterol (T-C) and the ratio of Total Cholesterol to HDL Cholesterol (T-C/HDL-C). Total cholesterol is a measure of all the cholesterol in your blood, while HDL cholesterol is often called 'good' cholesterol because it helps remove other forms of cholesterol from your bloodstream. The ratio between the two provides an indication of cardiovascular risk. It's like looking at both the total amount of something and the proportion that is considered beneficial.
• Underwriting categories: The table presents different underwriting categories, such as 'Elite Plus,' 'Preferred Plus,' and 'Standard Plus.' These categories likely represent different levels of risk as assessed by the insurance company, with 'Elite Plus' being the lowest risk and 'Standard' or 'Standard Plus' being higher risk. Each category has its own set of acceptable cholesterol ranges, which are probably stricter for the lower-risk categories.
• Table structure: The table is organized with rows representing different age groups (e.g., '54 and younger,' '55 to 69') and columns representing the different underwriting categories. Each cell in the table shows the maximum acceptable total cholesterol value and the corresponding T-C/HDL-C ratio for that age group and underwriting category. For example, for individuals aged 54 and younger in the 'Elite Plus' category, the maximum total cholesterol might be 220 mg/dL and the maximum T-C/HDL-C ratio might be 4.5.
• Implication for LDL-C: The reference text points out that LDL cholesterol, often called 'bad' cholesterol, is not used in these underwriting guidelines, while total cholesterol and the T-C/HDL-C ratio are used together. This suggests that, at least for the insurance company that created these guidelines, overall cholesterol levels and the balance between total and 'good' cholesterol are considered more important indicators of risk than LDL-C alone.

Scientific Validity

• Relevance to the study: This table provides an interesting perspective on how cholesterol is assessed in a real-world setting outside of clinical practice. It highlights the potential difference between clinical guidelines, which often emphasize LDL-C, and the criteria used by insurance companies, which may focus on other lipid measures. However, it is important to note that these are specific underwriting guidelines from one company and may not be representative of the entire insurance industry. The connection between life insurance underwriting practices and the relationship between LDL-C and mortality (the main focus of the study) may be considered tangential.
• Validity of underwriting criteria: The specific cholesterol ranges and ratios used in the table are based on the insurance company's internal risk assessment models. The scientific validity of these models is not directly assessed in the study. It is possible that these models are based on actuarial data and internal analyses, but the specific data and methods used are not provided.
• Emphasis on T-C and HDL-C: The emphasis on total cholesterol and the T-C/HDL-C ratio in the underwriting guidelines is noteworthy, given the current clinical focus on LDL-C. This could reflect a difference in the timeframes considered (long-term mortality risk for insurance vs. shorter-term cardiovascular risk in clinical settings) or a difference in the specific outcomes being assessed.
• Source of the guidelines: The footnote indicates that the source of the table is a specific life insurance company's underwriting guide. The generalizability of these guidelines to other insurance companies or to clinical practice is unknown. It would be helpful to know if these guidelines are based on industry standards or are specific to this particular company.

Communication

• Clarity of caption: The caption clearly states the content of the table, including the lipid measures presented (total cholesterol and T-C/HDL-C ratio) and the context (life insurance underwriting categories).
• Organization and layout: The table is well-organized, with clear headings for each column (underwriting categories) and row (age groups). The use of separate rows for different age groups allows for an assessment of how cholesterol criteria may vary with age. However, the specific layout and formatting details are not provided in the text and would need to be assessed directly from the table itself.
• Use of abbreviations: The table uses abbreviations (T-C, HDL-C). These abbreviations are clearly defined in the caption.
• Accessibility to non-experts: The table presents relatively straightforward data, but the concept of life insurance underwriting and the specific categories used may be unfamiliar to some readers. Providing a brief explanation of these concepts in the introduction or discussion section could enhance comprehension for a broader audience.
• Completeness of information: The footnote provides the source of the underwriting guidelines, which is important for transparency. However, it could be further improved by providing more details about the specific criteria used for each underwriting category, beyond just cholesterol levels. Additionally, explaining why certain age ranges were chosen would add clarity.

Methods

Key Aspects

• Study Design and Setting: This study employed a retrospective cohort design, analyzing data from the University of Pittsburgh Medical Center (UPMC) electronic medical record (EMR) system. The study period spanned from January 4, 2000, to December 31, 2022. This approach leverages a large, real-world dataset, enhancing the generalizability of the findings to routine clinical practice.
• Data Sources: The study utilized health-related data from the UPMC EMR and its associated clinical systems. These data were aggregated and standardized within a clinical data warehouse. The researchers accessed sociodemographic information, medical history, billing charges, diagnoses, and procedures, coded according to the International Classification of Diseases, 9th and 10th Revisions. Mortality data were obtained from hospital discharge records and the Social Security Administration's Death Master File. Cause of death was not available for this analysis. A composite outcome of ASCVD was also ascertained from UPMC records, defined as the occurrence of myocardial infarction, stroke, percutaneous coronary intervention, coronary artery bypass graft surgery, or peripheral vascular disease.
• Eligibility Criteria: The study focused on adults aged 50-89 years. The index date for patient selection was the first cholesterol measurement date. Patients were required to have non-missing values for total cholesterol (T-C), LDL-C, and HDL-C. The population was restricted to 'primary prevention' patients, defined as those without a history of diabetes, coronary artery disease, carotid artery disease, peripheral vascular disease, cardiac arrest, hemorrhagic or ischemic stroke, or transient ischemic attack. Participants were also required to be not on statin therapy at baseline or within one year of follow-up. To mitigate potential reverse causation, patients who died within one year of the baseline cholesterol measurement or had baseline T-C ≤120 mg/dL or LDL-C <30 mg/dL were excluded.
• Classification of Lipid Levels: Patients were categorized into mutually exclusive lipid-level groups based on common clinical thresholds. LDL-C categories were 30-79, 80-99, 100-129, 130-159, 160-189, and ≥190 mg/dL. T-C categories were 121-160, 161-200, 201-240, 241-280, and ≥281 mg/dL. The T-C/HDL-C ratio and triglycerides/HDL-C ratio were also analyzed. The LDL-C category of 80-99 mg/dL was selected as the referent group to mitigate potential bias due to reverse causation.
• Outcome Measures: The primary outcome was all-cause mortality, with follow-up starting 365 days after the baseline cholesterol measurement. For those who did not die, follow-up ended at their last EMR record. A secondary outcome was the occurrence of ASCVD.
• Statistical Analysis: Descriptive statistics (median, IQR, counts, percentages) were used to characterize the study population. The Kaplan-Meier method was used to estimate cumulative mortality rates. Cox regression was employed to estimate hazard ratios (HRs) for mortality, with a crude model and an adjusted model accounting for covariates and statin initiation after one year. Secondary analyses evaluated lipid parameters in relation to mortality risk using non-parametric generalized additive models with smoothing splines. Subgroup analyses were conducted based on age, sex, and baseline ASCVD risk classification.

Strengths

• Comprehensive Data Sources

  The study leverages a robust data infrastructure, integrating multiple clinical and administrative data sources within the UPMC system, enhancing data completeness and validity.

  "Health- related data captured in the UPMC EMR and its ancillary clinical systems were aggregated and harmonised in a clinical data warehouse" (Page 2)
• Clear Eligibility Criteria

  The study defines clear and specific eligibility criteria, ensuring a well-defined study population and reducing potential confounding.

  "The patient population was restricted to ‘primary prevention’ patients, defined as no history of diabetes" (Page 3)
• Appropriate Statistical Methods

  The use of Kaplan-Meier and Cox regression analyses is appropriate for time-to-event data, and the use of non-parametric models provides flexibility in exploring the relationship between lipid parameters and mortality.

  "For each LDL- C category, the Kaplan- Meier method was used to calculate cumulative mortality rates" (Page 3)
• Reverse Causation Mitigation

  The study design explicitly addresses potential reverse causation by excluding patients who died within one year of baseline and those with very low cholesterol levels.

  "to help offset potential bias due to reverse causation (ie, very low cholesterol being a marker for malnutrition and overall poor health)" (Page 3)

Suggestions for Improvement

• Clarify Covariate Selection

  This medium-impact improvement would enhance the transparency of the statistical analysis and allow readers to better understand the potential confounders considered. The Methods section is the appropriate place for this clarification as it provides the details of the study's analytical approach.

  "an adjusted model that included covariates selected by a forward stepwise approach" (Page 3)

  Implementation: Provide a list of the specific covariates included in the adjusted Cox regression model. For example: "The adjusted model included the following covariates: age, sex, race, body mass index, smoking status, history of hypertension, history of atrial fibrillation, history of heart failure, history of cancer, history of chronic obstructive pulmonary disease, history of chronic kidney disease, baseline systolic blood pressure, baseline diastolic blood pressure, baseline glucose level, use of ACE inhibitors, use of beta-blockers, use of calcium channel blockers, use of diuretics, use of aspirin, use of direct oral anticoagulants, use of antidepressants, use of opioids, and statin initiation after one year of follow-up."

• Elaborate on Handling of Missing Data

  This low-impact improvement would provide a more complete picture of the data quality and potential limitations. While the requirement for non-missing lipid values is mentioned, a brief discussion of how other missing data were handled would further strengthen the methodological rigor.

  "For analysis, we required non- missing laboratory values for T- C, LDL- C and HDL- C." (Page 2)

  Implementation: Add a sentence or two describing the approach to missing data for covariates other than the primary lipid measurements. For example: "Patients with missing data for any of the covariates included in the adjusted model were excluded from that specific analysis. Sensitivity analyses were conducted to assess the potential impact of missing data on the main findings."

• Justify Choice of Referent Group

  This medium-impact improvement would provide a clearer rationale for the selection of the LDL-C 80-99 mg/dL category as the referent group. While the reason is mentioned (to mitigate potential bias), further elaborating on this choice would enhance the reader's understanding of the study's design.

  "we selected the LDL- C category of 80- 99mg/dL as the referent group, rather than the lowest LDL- C group (30- 79mg/dL)." (Page 3)

  Implementation: Expand the explanation for choosing the LDL-C 80-99 mg/dL category as the referent group. For example: "The LDL-C category of 80-99 mg/dL was chosen as the referent group because it represents a commonly observed range in the general population and is less likely to be influenced by underlying health conditions that might lead to very low LDL-C levels. This approach helps to minimize potential bias due to reverse causation, where low LDL-C could be a marker of poor health rather than a cause of increased mortality."

Results

Key Aspects

• Patient Characteristics by LDL-C Category: The study population was categorized into six groups based on baseline LDL-C levels: 30-79, 80-99, 100-129, 130-159, 160-189, and 190 mg/dL or higher. The mean LDL-C was 119 mg/dL, and the median age was 59 years. Notably, the group with the lowest LDL-C (30-79 mg/dL) exhibited a higher prevalence of comorbidities and risk factors, suggesting potential reverse causation, where underlying illness might contribute to both low LDL-C and increased mortality. Conversely, the highest LDL-C category (≥190 mg/dL) had the highest estimated 10-year ASCVD risk.
• Patient Follow-Up and Mortality Rates: Patients were followed for a mean of 6.1 years, with 17% having 10 or more years of follow-up. Across the LDL-C categories, the mean follow-up among those who did not die ranged from 5.8 to 6.4 years. The study found a U-shaped relationship between LDL-C categories and 10-year cumulative mortality rates. The lowest mortality rates were observed in the LDL-C range of 100-189 mg/dL (11.7%, 10.7%, and 10.1% for the 100-129, 130-159, and 160-189 mg/dL groups, respectively), while the highest rates were seen in the lowest (30-79 mg/dL, 19.8%) and highest (≥190 mg/dL, 14.0%) LDL-C categories.
• Adjusted Hazard Ratios for Mortality: After adjusting for covariates, the three LDL-C categories within the range of 100-189 mg/dL showed a significantly lower risk of mortality compared to the referent group (LDL-C 80-99 mg/dL). The adjusted hazard ratios (HRs) were 0.87 (95% CI 0.83-0.91) for 100-129 mg/dL, 0.88 (95% CI 0.84-0.93) for 130-159 mg/dL, and 0.91 (95% CI 0.84-0.98) for 160-189 mg/dL. In contrast, the lowest LDL-C category (30-79 mg/dL) had a significantly higher mortality risk (HR 1.23, 95% CI 1.17-1.30), and the highest category (≥190 mg/dL) also showed an increased risk (HR 1.19, 95% CI 1.06-1.34).
• Assessment of ASCVD: Similar to the mortality findings, a U-shaped relationship was observed between LDL-C categories and 10-year cumulative rates of ASCVD. The lowest rates were found in the 100-189 mg/dL range. Adjusted HRs for ASCVD risk in this range were similar to or nominally lower than the referent group. Baseline ASCVD risk categories (low, medium, high) were strongly associated with 10-year ASCVD rates.
• Subgroup Analyses: Subgroup analyses by age (50-69 and 70-89 years) and sex (female and male) showed that the three LDL-C categories within the 100-189 mg/dL range consistently exhibited similar or slightly lower mortality risk compared to the referent group. When stratified by 10-year ASCVD risk score, the 100-189 mg/dL categories again showed relatively similar and statistically lower mortality risk compared to the referent group.
• Secondary Lipid Measures: The study also evaluated secondary lipid measures, including the total cholesterol (T-C)/HDL-C ratio and the triglycerides/HDL-C ratio. Patients with a T-C/HDL-C ratio >6.0 had a significantly higher mortality risk than those with a ratio ≤3.0. The triglycerides/HDL-C ratio showed a gradient relationship with mortality, with lower values (quintiles) conferring lower mortality risk. These secondary lipid measures appeared to be more predictive of mortality than LDL-C alone.
• Evaluation of Potential Reverse Causation: To address potential reverse causation, the study excluded deaths that occurred within the first year after baseline LDL-C measurement. Analysis of these excluded deaths revealed a higher proportion in the lowest LDL-C category (30-79 mg/dL) compared to its prevalence in the primary analysis population, supporting the decision to exclude these early deaths and the choice of the 80-99 mg/dL LDL-C category as the referent group.

Strengths

• Clear Presentation of Baseline Characteristics

  The Results section effectively presents the baseline characteristics of the study population, stratified by LDL-C categories, providing a clear understanding of the cohort's composition and potential confounders.

  "The mean (SD) LDL-C was 119 (31) mg/dL, and the prevalence of patients within the six LDL-C categories was as follows: 30–79 (9.1%), 80–99 (18.3%), 100–129" (Page 3)
• Comprehensive Reporting of Mortality Outcomes

  The section thoroughly reports the primary outcome of all-cause mortality, including cumulative mortality rates, adjusted hazard ratios, and subgroup analyses, providing a comprehensive picture of the relationship between LDL-C and mortality.

  "Adjusted mortality HRs and 95% CIs (table 2), as compared with the referent group of LDL-C 80-99 mg/dL, were as follows: 30–79 mg/dL (1.23, 95% CI" (Page 6)
• Inclusion of Secondary Outcomes and Analyses

  The inclusion of secondary outcomes, such as ASCVD, and analyses of secondary lipid measures adds depth to the study and provides a more holistic view of the relationship between lipid profiles and cardiovascular health.

  "Assessment of ASCVD In ascending order from lowest LDL-C category (30-79mg/dL) to highest LDL-C category (≥190mg/" (Page 6)
• Addressing Potential Reverse Causation

  The study explicitly addresses the potential for reverse causation by excluding early deaths and providing a rationale for the choice of the referent LDL-C group, demonstrating a thoughtful approach to study design and analysis.

  "Evaluation of potential reverse causation By study design, the 2494 patient deaths that occurred from baseline LDL-C measurement to 365 days were excluded from the primary analysis." (Page 7)

Suggestions for Improvement

• Enhance Clarity of Statistical Adjustments

  This medium-impact improvement would enhance the transparency and interpretability of the results. While the Methods section mentions covariate adjustment, explicitly stating the adjusted covariates in the Results section when presenting the hazard ratios would allow readers to immediately understand the factors accounted for in the analysis.

  "Adjusted mortality HRs and 95% CIs (table 2), as compared with the referent group of LDL-C 80-99 mg/dL" (Page 6)

  Implementation: When presenting the adjusted hazard ratios in the Results section (e.g., in Table 2's footnote or within the text), list the specific covariates included in the adjustment model. For example: "Adjusted hazard ratios were calculated after controlling for age, sex, race, BMI, smoking status, history of hypertension, history of diabetes, and use of antihypertensive medications."

• Provide More Context for ASCVD Results

  This low-impact improvement would provide a more complete understanding of the ASCVD findings. While the Results section presents the U-shaped relationship between LDL-C and ASCVD, briefly discussing the clinical significance of these findings and how they relate to the primary mortality outcome would enhance the reader's interpretation.

  "Assessment of ASCVD In ascending order from lowest LDL-C category (30-79mg/dL) to highest LDL-C category (≥190mg/" (Page 6)

  Implementation: Add a sentence or two after presenting the ASCVD results, explaining their clinical implications. For example: "Although a U-shaped relationship was observed between LDL-C and ASCVD, the absolute differences in ASCVD rates across the LDL-C categories were relatively small. This suggests that while LDL-C may play a role in ASCVD development, other factors likely contribute significantly to overall cardiovascular risk in this primary prevention population. These findings further support the primary mortality results, highlighting the limited utility of LDL-C as an isolated predictor of adverse outcomes."

• Clarify the Rationale for Referent Group Selection in Results

  This medium-impact improvement would strengthen the Results section by providing a clear justification for the choice of the referent group within the section itself. While the Methods section explains this choice, reiterating the rationale briefly in the Results section would enhance the reader's understanding of the presented comparisons and reinforce the study's methodological rigor.

  "Adjusted mortality HRs and 95% CIs (table 2), as compared with the referent group of LDL-C 80-99 mg/dL" (Page 6)

  Implementation: When first introducing the referent group in the Results section, add a brief phrase explaining the rationale. For example: "As described in the Methods, the LDL-C category of 80-99 mg/dL was selected as the referent group to mitigate potential bias due to reverse causation. Therefore, adjusted mortality HRs are presented in comparison to this group."

Non-Text Elements

Figure 1 Plot of cumulative mortality rates in 6-month intervals over 12 years...

Full Caption

Figure 1 Plot of cumulative mortality rates in 6-month intervals over 12 years of follow-up by baseline low-density lipoprotein cholesterol (LDL-C) category. Dashed lines depict the three lowest LDL-C categories (30–79, 80–99 and 100–129mg/dL) and solid lines depict the highest LDL-C categories (130–159, 160–189 and ≥190 mg/dL).

Figure/Table Image (Page 7)

First Reference in Text

In ascending order from lowest LDL-C category (30-79 mg/dL) to highest LDL-C category (≥190mg/ dL), 10-year cumulative mortality rates were U-shaped at 19.8%, 14.7%, 11.7%, 10.7%, 10.1% and 14.0% (table 2, figures 1 and 2).

Description

• Type of plot: This is a line graph, a visual representation where data points are connected by lines to show trends over time.
• X-axis: The horizontal axis (x-axis) represents time, spanning 12 years and divided into 6-month intervals. This means each tick mark on the axis corresponds to a half-year period, allowing us to see how something changes over the course of a dozen years.
• Y-axis: The vertical axis (y-axis) represents the cumulative mortality rate. 'Cumulative' means the total percentage of people who have died up to a given point in time. So, as you move up the y-axis, the percentage of deaths increases.
• Data groups: The graph shows six different lines, each representing a group of people categorized by their baseline LDL-C levels. LDL-C stands for low-density lipoprotein cholesterol, often referred to as 'bad' cholesterol because high levels are associated with an increased risk of heart disease. The categories range from low (30-79 mg/dL) to very high (≥190 mg/dL), with 'mg/dL' meaning milligrams per deciliter, a unit used to measure the concentration of a substance in a liquid (in this case, cholesterol in blood).
• Line styles: The lines are differentiated by style: dashed lines for the three lowest LDL-C categories and solid lines for the three highest. This visual distinction helps to quickly compare the trends between groups with lower versus higher initial cholesterol levels.
• Overall trend: The graph illustrates how the percentage of deaths within each LDL-C category changes over the 12-year period. By following each line, we can see whether mortality increases steadily, levels off, or shows other patterns over time.

Scientific Validity

• Relevance of LDL-C categorization: The categorization of LDL-C is consistent with clinical guidelines, although the specific cut-off points might be debated. The use of standard categories allows for comparison with other studies and clinical practice.
• Suitability of cumulative mortality rate: Cumulative mortality rate is an appropriate measure for assessing the long-term risk associated with different LDL-C levels. It reflects the overall impact of LDL-C on survival over the study period.
• Follow-up duration: A 12-year follow-up is substantial for this type of study and allows for the observation of long-term trends. However, it's important to note that even longer follow-up periods might reveal different patterns.
• Clarity of methodology in figure: While the figure itself does not detail the methodology used to collect or analyze the data, it is implied that standard survival analysis techniques were employed. The reference text confirms this, and the methods section of the paper should provide further details. The specific statistical methods used to generate the cumulative mortality rates are not detailed in the figure or the provided reference text, which could be a limitation for readers trying to replicate or critically evaluate the analysis.

Communication

• Visual clarity: The use of different line styles (dashed vs. solid) for different LDL-C categories is effective in visually distinguishing between the groups. However, the specific colors used for each line are not described in the provided information and should be clearly indicated in the figure legend for optimal clarity.
• Labeling: The axes are labeled with appropriate units (years for the x-axis and presumably percentage for the y-axis, although the specific unit is not stated in the caption). The LDL-C categories are clearly labeled in the caption.
• Caption completeness: The caption provides a concise description of the figure's content, including the meaning of the different line styles and the LDL-C categories. However, it could be improved by explicitly stating the unit for the y-axis (e.g., 'cumulative mortality rate (%)').
• Interpretation guidance: The caption does not provide any interpretation of the observed trends, leaving that to the results section of the paper. This is appropriate for a figure caption, which should primarily describe what is shown rather than interpret its meaning.

Figure 2 Plot of mortality HRs (filled circles) and 95% Cls (vertical lines)...

Full Caption

Figure 2 Plot of mortality HRs (filled circles) and 95% Cls (vertical lines) across categories of LDL cholesterol (top), total cholesterol to HDL cholesterol ratio (middle) and triglycerides to HDL cholesterol ratio (bottom). The left side of the graph is for patients aged 50-69 years; the right side is for patients aged 70-89 years. The dashed line reflects the referent group null value (1.0) for the HR. Q: quintile. Each model is adjusted for age, race, sex, BMI, current smoker, former smoker and history of the following in the past year: hypertension, atrial fibrillation, arrhythmia, congestive heart failure, cancer, chronic obstructive pulmonary disease, chronic kidney disease, baseline systolic and diastolic blood pressure, glucose and the following medications in the past year: ACE inhibitors, beta-blockers, calcium blockers, any SBP lowering medication, diuretics, aspirin, DOACs, antidepressants, opioids and statin initiation >1 year after baseline cholesterol measurement. BMI, Body Mass Index; DOAC, direct oral anticoagulant; HDL, high-density lipoprotein; LDL, low-density lipoprotein; SBP, systolic blood pressure.

Figure/Table Image (Page 8)

First Reference in Text

For the two different age groups, the three LDL-C catego- ries within the range of 100–189mg/dL showed relatively similar and slightly lower mortality risk compared with the referent group of LDL-C 80–99 mg/dL (table 2, figure 2).

Description

• Overall structure: This figure is a complex graph divided into six sub-graphs arranged in a 2x3 grid. It compares the risk of death across different categories of cholesterol levels and ratios, separated by age group.
• Data presentation: Each sub-graph uses a dot-and-whisker plot. The 'dot' (filled circle) represents the Hazard Ratio (HR), which is a measure of how often a particular event (in this case, death) happens in one group compared to another group over a certain period. The 'whiskers' (vertical lines) represent the 95% Confidence Interval (CI) for each HR. The 95% CI gives a range of values within which we are 95% confident that the true HR lies.
• X-axis of sub-graphs: The horizontal axis (x-axis) of each sub-graph represents different categories of cholesterol levels or ratios. For example, in the top sub-graphs, it shows categories of LDL cholesterol (often called 'bad' cholesterol). The middle sub-graphs show categories of the ratio of total cholesterol to HDL cholesterol ('good' cholesterol), and the bottom sub-graphs show categories of the ratio of triglycerides (another type of fat in the blood) to HDL cholesterol.
• Y-axis of sub-graphs: The vertical axis (y-axis) represents the Hazard Ratio (HR). A dashed horizontal line is drawn at HR = 1.0. This line represents the 'null value,' meaning no difference in risk between the groups being compared. If a dot is above this line, it suggests a higher risk of death in that category compared to the reference group. If a dot is below the line, it suggests a lower risk.
• Age groups: The left side of the entire graph shows data for patients aged 50-69 years, while the right side shows data for patients aged 70-89 years. This allows for a comparison of how cholesterol levels and ratios affect mortality risk in different age groups.
• Statistical adjustments: The caption states that each 'model' (each sub-graph) is 'adjusted' for a long list of factors, including age, race, sex, Body Mass Index (BMI), smoking status, various health conditions, and medication use. This means that the researchers have used statistical methods to account for the potential influence of these factors on the relationship between cholesterol and mortality. This is important because these factors could independently affect the risk of death.
• Quintiles: The abbreviation 'Q' in the caption likely refers to 'quintile.' This means that for the ratios (total cholesterol to HDL and triglycerides to HDL), the data has been divided into five equal groups based on the distribution of these ratios in the study population.

Scientific Validity

• Appropriateness of Hazard Ratios: Hazard ratios are an appropriate measure for comparing mortality risk between groups in a cohort study. They provide a relative measure of risk, which is useful for understanding the impact of different cholesterol levels.
• Validity of statistical adjustments: The extensive list of covariates adjusted for in the analysis is commendable. This helps to isolate the specific effect of cholesterol on mortality, reducing the risk of confounding. However, the specific methods used for adjustment are not described in the caption and should be detailed in the methods section.
• Choice of reference group: The choice of the 80-99 mg/dL LDL-C category as the reference group is justified in the paper as a group with putatively lower risk than the lowest category, but may still be debated. The rationale for this choice should be clearly explained in the methods section.
• Generalizability: The generalizability of the findings may be limited by the specific characteristics of the study population (e.g., patients from a particular healthcare system). The demographics and clinical characteristics of the cohort should be thoroughly described to allow readers to assess the applicability of the results to other populations.

Communication

• Complexity: The figure is highly complex, presenting a large amount of information in a condensed format. This may make it challenging for readers to fully grasp the findings without careful study.
• Clarity of labeling: The labeling of the axes and categories within each sub-graph is not described in the provided caption, but should be clear and consistent across all sub-graphs. The use of abbreviations (BMI, DOAC, HDL, LDL, SBP) is acceptable given that they are defined in the caption.
• Use of visual cues: The use of filled circles for HRs and vertical lines for 95% CIs is standard and effective. The dashed line at HR = 1.0 is a helpful visual aid for interpreting the results.
• Caption completeness: The caption provides a detailed description of the figure's content, including the variables plotted, the age groups represented, and the statistical adjustments made. However, it could be improved by explicitly stating what is plotted on the x-axis of each subgraph (e.g., categories of LDL cholesterol).
• Interpretation guidance: The caption does not provide any interpretation of the observed trends, leaving that to the results section. This is appropriate for a figure caption.

Table 1 Baseline characteristics of study population by baseline LDL...

Full Caption

Table 1 Baseline characteristics of study population by baseline LDL cholesterol value

Figure/Table Image (Page 4)

First Reference in Text

The median age of patients was 59 years and mean age ranged nominally across the six LDL-C categories from 60.7 to 61.7 years. There was a general indication of overall higher baseline risk in the group of patients with LDL-C from 30 to 79mg/dL (table 1) (consistent with the stated concern of potential reverse causation).

Description

• Table purpose: This table presents the characteristics of the study participants at the beginning of the study, categorized by their baseline LDL cholesterol values. Think of it like a detailed description of the different groups of people being studied, before any interventions or treatments are applied. This is important to understand the composition of each group and whether they differ in ways other than their cholesterol levels.
• Table structure: The table is organized with rows representing different characteristics of the participants (like age, sex, race, medical history, etc.), and columns representing different categories of LDL cholesterol levels. LDL cholesterol, or low-density lipoprotein cholesterol, is often referred to as 'bad' cholesterol. These categories likely range from low to high, allowing for a comparison of characteristics across different levels of LDL cholesterol.
• Data presented: For each characteristic, the table shows either the median and interquartile range (IQR) for continuous variables (like age or blood pressure) or counts and percentages for categorical variables (like sex or history of a certain disease). The median is the middle value when all values are arranged in order, and the IQR gives an idea of the spread of the data around the median. Counts and percentages show how many participants in each LDL-C category have a particular characteristic.
• Baseline characteristics: The characteristics listed in the table are called 'baseline characteristics' because they describe the participants at the start of the study. These include demographic information (age, sex, race), medical history (e.g., history of obesity, hypertension, diabetes), and lifestyle factors (e.g., smoking status). They provide a snapshot of the health status and other relevant factors of the participants before any changes that might occur during the study.
• Importance of baseline data: Comparing baseline characteristics across different LDL-C categories is crucial for several reasons. First, it helps to identify any potential confounding factors, which are factors other than LDL-C that might influence the outcome of the study. For example, if one LDL-C group has a much higher percentage of smokers, this could affect their risk of death independently of their cholesterol levels. Second, it allows researchers to assess whether the groups are comparable at the start of the study, which is important for drawing valid conclusions about the effects of LDL-C.

Scientific Validity

• Comprehensiveness of characteristics: The table appears to include a comprehensive list of relevant demographic, clinical, and lifestyle characteristics. This allows for a thorough assessment of potential confounders and baseline differences between the LDL-C groups.
• Appropriateness of statistical measures: The use of median and IQR for continuous variables and counts/percentages for categorical variables is appropriate for describing the distribution of baseline characteristics.
• Potential for residual confounding: While the table includes many important characteristics, there is always a possibility of residual confounding due to unmeasured or unknown factors. The authors should acknowledge this limitation in the discussion section.
• Relevance to reverse causation: The reference text highlights a higher baseline risk in the lowest LDL-C group, which could be indicative of reverse causation (i.e., underlying illness leading to low LDL-C rather than low LDL-C causing illness). The table provides the necessary data to further investigate this possibility by examining the prevalence of specific illnesses in this group.

Communication

• Clarity of caption: The caption clearly states the purpose and content of the table.
• Organization and layout: The table is well-organized, with clear headings for each column (LDL-C categories) and row (characteristic). The use of bold font for subheadings within the table enhances readability. However, the specific layout and formatting details are not provided in the text and would need to be assessed directly from the table itself.
• Use of abbreviations: The table likely uses abbreviations for some characteristics (e.g., BMI, HDL, LDL). These abbreviations should be clearly defined in a footnote or legend accompanying the table.
• Accessibility to non-experts: While the table presents complex data, the clear labeling and organization make it relatively accessible to readers with some background knowledge. However, a brief explanation of the importance of baseline characteristics and the concept of confounding in the results or discussion section could further enhance understanding for a broader audience.