Association of Obstructive Sleep Apnea With Risk of Myocardial Infarction: A Multivariable Mendelian Randomization and Mediation Analysis

This study utilizes multivariable Mendelian randomization to establish a causal link between obstructive sleep apnea and myocardial infarction, revealing that the association is partially mediated by obesity and atrial fibrillation rather than solely by traditional hemodynamic factors.

The Gist

The Big Question: Is Sleep Apnea the "Arsonist" or Just a "Witness"?

Imagine your heart is a high-performance car. Myocardial Infarction (MI) is a major engine failure (a heart attack). Obstructive Sleep Apnea (OSA) is a condition where your throat collapses while you sleep, causing you to stop breathing repeatedly.

For years, doctors have noticed that people with sleep apnea often get heart attacks. But they weren't sure who was causing whom.

• The Old Theory: Maybe sleep apnea is the arsonist setting the fire (causing the heart attack).
• The Skeptic's View: Or maybe sleep apnea is just a witness standing next to the fire, because both the arsonist and the witness are caused by a third thing, like being overweight (obesity) or smoking.

This study wanted to find out: Is sleep apnea actually causing the heart attacks, or is it just hanging out with the other bad guys?

The Detective Tool: "Genetic Time Travel"

To solve this, the researchers used a method called Mendelian Randomization (MR). Think of this as genetic time travel.

Instead of asking people, "Do you sleep well?" and "Do you have heart problems?" (which is messy because people lie, forget, or have other habits), the researchers looked at people's DNA.

• Your DNA is like a lottery ticket you get at birth. It's random and unchangeable.
• If you have certain "sleep apnea tickets" in your DNA, you are genetically destined to have a higher risk of sleep apnea, regardless of your diet or lifestyle.
• By tracking these "tickets," the researchers could see if people with the "sleep apnea tickets" also got "heart attack tickets" more often.

Because your DNA is set before you are born, it rules out the idea that the heart attack caused the sleep apnea (reverse causality) or that a bad lifestyle caused both. It isolates the relationship between the two.

The Investigation: What Did They Find?

The researchers acted like detectives, running the experiment in three different ways to make sure they didn't make a mistake.

1. The Direct Link (The Smoking Gun)
They found that people genetically prone to sleep apnea did have a higher risk of heart attacks.

• The Verdict: Sleep apnea isn't just a witness; it is likely a co-conspirator in causing heart attacks. It has a direct, causal link.

2. The "Who Did It?" Investigation (Mediation Analysis)
Okay, so sleep apnea causes heart attacks. But how? Does it do it alone, or does it hire other bad guys? The researchers checked 13 potential "hired guns," including obesity, high blood pressure, and diabetes.

• The Obesity Connection: They found that Body Mass Index (BMI) is a huge player. About 36% of the damage sleep apnea does to the heart is because sleep apnea makes you gain weight (or vice versa), and that extra weight hurts the heart. It's like sleep apnea handing a heavy backpack to the heart, and the backpack is the real killer.
• The Blood Pressure Surprise: Surprisingly, High Blood Pressure (Systolic BP) wasn't the main culprit in this specific chain. While high blood pressure is bad for the heart generally, it didn't explain why sleep apnea leads to heart attacks as much as obesity did.
• The Hidden Villain (Atrial Fibrillation): They found a very interesting clue involving Atrial Fibrillation (AF), which is a chaotic, fluttering heart rhythm. Sleep apnea seems to scramble the heart's electrical wiring (causing AF), and that scrambling leads to heart attacks. This happens even if your blood pressure is normal.

3. The "Head-to-Head" Showdown (Multivariable MR)
The researchers put all these factors in a ring to see who was the strongest fighter.

• When they adjusted for Obesity and Blood Pressure, the direct link between sleep apnea and heart attacks got weaker, but it didn't disappear.
• However, when they adjusted for Atrial Fibrillation, the link between sleep apnea and heart attacks became very shaky. This suggests that sleep apnea might be causing heart attacks by first causing heart rhythm problems.

The Reverse Check

They also asked: "Does having a heart attack cause sleep apnea?"

• The Answer: No. The genetic evidence showed that heart attacks do not cause sleep apnea. The arrow of causality only points one way: Sleep Apnea $\rightarrow$ Heart Attack.

The Takeaway: What Does This Mean for You?

Think of your body as a house.

• Sleep Apnea is a leaky roof.
• Heart Attack is the house burning down.

This study tells us that the leaky roof (sleep apnea) is indeed causing the fire. But it's not just the water from the roof; the water is ruining the foundation (Obesity) and short-circuiting the electrical system (Atrial Fibrillation), which eventually burns the house down.

The Clinical Lesson:
If you have sleep apnea, treating it isn't just about getting a better night's sleep. It's about fire prevention.

1. Weight Management: Since obesity is a major bridge between sleep apnea and heart attacks, losing weight might break the chain.
2. Heart Rhythm Checks: Doctors should pay extra attention to heart rhythm issues (like AF) in sleep apnea patients, because the heart's electrical system might be getting scrambled by the lack of oxygen at night.
3. Don't Ignore the Sleep: Treating sleep apnea (like with a CPAP machine) might be a powerful way to protect your heart, not just your lungs.

In short: Sleep apnea is a silent, dangerous arsonist for your heart, and it works by making you heavier and scrambling your heart's electrical wiring. Fixing the sleep issue could save your heart.

Technical Summary

1. Problem Statement

Observational studies have consistently reported an association between Obstructive Sleep Apnea (OSA) and Myocardial Infarction (MI). However, these studies are limited by residual confounding (e.g., shared risk factors like obesity, smoking, and hypertension) and reverse causality (MI potentially influencing sleep patterns). It remains uncertain whether OSA is a direct causal risk factor for MI or if the association is merely a byproduct of these shared comorbidities. This study aims to establish a causal link using genetic epidemiology and to dissect the specific biological pathways (mediators) through which OSA might influence MI risk.

2. Methodology

The study employed a Two-Sample Mendelian Randomization (MR) framework using large-scale Genome-Wide Association Study (GWAS) summary statistics.

• Data Sources:
  • Exposure (OSA): Primary data from FinnGen (8,998 cases, 356,128 controls). Validation used alternative FinnGen definitions and an independent dataset (EBI).
  • Outcome (MI): Primary data from UK Biobank (14,394 cases, 361,194 controls). Validation used CARDIoGRAMplusC4D consortium data.
  • Mediators: 13 potential mediators including obesity traits (BMI, WHR), glycemic traits (T2D, HbA1c), lipids (LDL, HDL, TG), blood pressure (SBP), atrial fibrillation (AF), and lifestyle factors (smoking, alcohol).
• Instrumental Variable (IV) Selection:
  • SNPs associated with OSA were selected at a relaxed threshold ($P < 1 \times 10^{-5}$) to ensure sufficient statistical power, resulting in 54 independent SNPs (Mean F-statistic = 23.19).
  • Strict quality control was applied: LD clumping ($r^2 < 0.001$), removal of weak instruments ($F > 10$), Steiger filtering for directionality, and exclusion of variants with known pleiotropic associations with cardiometabolic traits.
• Analytical Framework:
  1. Primary MR: Inverse-Variance Weighted (IVW) method as the primary estimator, supplemented by Weighted Median and MR-Egger regression.
  2. Sensitivity Analyses: MR-PRESSO (outlier detection), Cochran's Q (heterogeneity), MR-Egger intercept (horizontal pleiotropy), and Leave-One-Out analysis.
  3. Mediation Analysis: Two-step MR to quantify the indirect effect of OSA on MI through 13 candidate mediators.
  4. Multivariable MR (MVMR): A progressive 6-step framework to estimate the direct effect of OSA on MI while simultaneously adjusting for correlated risk factors:
     • Model 1: BMI adjustment.
     • Model 2: BMI + Waist-to-Hip Ratio (WHR).
     • Model 3: BMI + Systolic Blood Pressure (SBP).
     • Model 4: Metabolic network (T2D, LDL, TG).
     • Model 5: Smoking + Atrial Fibrillation (AF).
     • Model 6: Head-to-head adjustment of AF and SBP.
  5. Reverse MR: To test if genetic liability to MI influences OSA risk.

3. Key Results

A. Causal Association (Primary MR)

• Genetically predicted OSA liability is significantly associated with an increased risk of MI.
• Odds Ratio (OR): 1.0024 per log-OR increase (95% CI: 1.0010–1.0039; $P=0.001$).
• Robustness: Results were consistent across different phenotype definitions and independent cohorts (CARDIoGRAMplusC4D). No significant heterogeneity or directional horizontal pleiotropy was detected.
• Reverse Causality: No evidence was found that genetic liability to MI influences OSA risk ($P=0.482$).

B. Mediation Analysis

• Body Mass Index (BMI): Identified as the most prominent mediator.
  • Indirect effect: 35.94% of the total OSA-MI association is mediated through BMI ($P=0.030$).
  • Note: This significance did not hold after False Discovery Rate (FDR) correction, suggesting it is an exploratory finding.
• Other Mediators: Systolic Blood Pressure (SBP) showed minimal mediation (0.28%; $P=0.678$). Type 2 Diabetes, lipids, and other factors showed no significant indirect effects.
• Bidirectional Link: Reverse MR confirmed a strong causal effect of BMI on OSA, indicating a bidirectional relationship.

C. Multivariable MR (MVMR) Findings

The progressive adjustment models revealed how the direct effect of OSA changes when accounting for specific pathways:

• Obesity: Adjusting for BMI attenuated the OSA effect but it remained significant. When WHR was added, the BMI effect vanished, but WHR (central adiposity) emerged as a significant independent risk factor, suggesting central fat distribution is more critical than general BMI in this pathway.
• Blood Pressure: In Model 3 (adjusting for BMI and SBP), SBP showed a strong independent effect on MI, but the direct effect of OSA became non-significant ($P=0.156$).
• Metabolic Network: Even after adjusting for T2D and lipids, OSA retained a significant direct effect ($P=0.015$).
• Arrhythmia (AF) vs. Hemodynamics: In the final "head-to-head" model (Model 6) adjusting for both AF and SBP:
  • AF remained a robust independent risk factor ($P=0.004$).
  • OSA retained only a marginal direct effect ($P=0.050$).
  • This suggests the OSA-MI link operates significantly through an arrhythmia-related pathway (OSA $\to$ AF $\to$ MI) that is distinct from hemodynamic load (SBP).

4. Key Contributions

1. Causal Validation: Provides the first robust genetic evidence supporting a causal relationship between OSA and MI, overcoming the limitations of observational confounding.
2. Pathway Dissection: Moves beyond a simple exposure-outcome association to map a multidimensional risk network. It identifies that the OSA-MI link is not solely driven by traditional hemodynamic factors (SBP) but involves complex interactions with central adiposity and electrophysiological remodeling (AF).
3. Mediator Differentiation: Highlights that while obesity is a major mediator, Atrial Fibrillation (AF) acts as a critical downstream node that may explain residual risk independent of blood pressure.
4. Methodological Rigor: Utilizes a progressive MVMR framework and extensive sensitivity analyses (including external validation cohorts) to ensure the findings are not artifacts of pleiotropy or population stratification.

5. Significance and Clinical Implications

• Risk Stratification: The findings suggest that cardiovascular risk assessment in OSA patients should extend beyond airway management. Clinicians should prioritize screening for central adiposity and atrial fibrillation.
• Mechanistic Insight: The study challenges the notion that blood pressure is the primary driver of OSA-related MI risk. Instead, it posits that intermittent hypoxia may drive cardiac remodeling and arrhythmogenesis (AF), which subsequently increases MI risk.
• Intervention Targets: While weight management is crucial (due to the BMI mediation), the identification of AF as a key intermediate suggests that early detection and management of arrhythmias in OSA patients could be a vital strategy for preventing MI.
• Limitations: The study relies on European ancestry data (limiting generalizability) and binary OSA definitions (lacking severity gradients). Some mediation results (like BMI) were exploratory and lost significance after FDR correction.

Conclusion: The study establishes OSA as a causal risk factor for MI, mediated partly by obesity but significantly driven by a pathway involving atrial fibrillation, independent of traditional blood pressure mechanisms. This supports a more holistic, multidimensional approach to managing cardiovascular risk in patients with sleep apnea.