Aspirin and outcomes in MASLD

Metabolic dysfunction-associated steatotic liver disease frequently coexists with cardiometabolic conditions, and many affected adults receive long-term antithrombotic therapy. Within this context, interest has focused on whether aspirin, a cornerstone of antiplatelet therapy, might alter hepatic outcomes. The recent multi-institutional analysis spanning three years reported associations between aspirin use alone and risks of all-cause mortality and liver-related events among people with metabolic dysfunction-associated steatotic liver disease. The report is observational and did not test a treatment strategy, but it offers a timely look at outcome patterns in routine care. As with any exposure-outcome analysis in real-world data, interpretation hinges on how cohorts, covariates, and endpoints were defined and analyzed.

The biologic rationale for potential hepatic effects of aspirin is multifaceted. Aspirin irreversibly inhibits cyclooxygenase-1, reducing thromboxane A2 and thereby platelet activation and aggregation, a pathway implicated in sinusoidal microthrombosis and inflammation that can contribute to fibrosis progression. Beyond platelet activation, low-dose aspirin can modulate proinflammatory eicosanoids and potentially influence hepatocellular injury signaling. Whether these mechanisms translate to measurable differences in decompensation, cirrhosis progression, or hepatocellular carcinoma incidence remains uncertain and likely depends on baseline risk, dose, adherence, and competing causes of morbidity and mortality. The present report adds clinical context by mapping exposure to outcomes over a defined follow-up.

Clinical decision making, however, must weigh potential benefits against risks, especially gastrointestinal and portal hypertensive bleeding. In people with advanced fibrosis or clinically significant portal hypertension, the hemorrhagic hazard from antiplatelet effects can offset any favorable hepatic signals. Real-world associations therefore reflect not only pharmacology but also prescribing patterns, co-therapies such as proton pump inhibitors, and careful selection by clinicians who may avoid aspirin in patients perceived at high bleeding risk. Recognizing these countervailing forces is essential when interpreting any observed differences in mortality or liver-related event rates.

The publication is accessible through PubMed and details a multi-institutional data set and three-year observation window, providing a contemporary view across diverse care settings. Readers can review the abstract and indexing at PubMed. While the report communicates associations rather than causal effects, it offers an opportunity to examine how aspirin exposure in routine practice aligns with hepatic outcomes that matter to patients and systems. The remainder of this article summarizes scope, endpoints, and methodological considerations relevant to critical appraisal.

Cohort identification and endpoints

The analysis centers on adults with documentation consistent with MASLD in real-world records aggregated across multiple institutions. Typically, such cohorts are identified using diagnostic codes, laboratory and imaging patterns, or noninvasive fibrosis indices. Endpoints for the hepatic domain commonly include decompensation events, incident or worsening ascites, variceal bleeding, hepatic encephalopathy, transplant referral or listing, development of hepatocellular carcinoma, and liver-related hospitalizations. Mortality is usually captured through linkage to administrative sources or health system records to estimate all-cause outcomes over fixed intervals. Exact definitions and adjudication processes govern the reliability of event capture and may vary across contributing sites.

Exposure categorization is critical. Aspirin use alone in routine care is often inferred from medication orders, fills, or structured documentation, with dose and adherence imperfectly measured. Time at risk is commonly defined from a qualifying baseline to outcome occurrence or censoring, and the analytic approach determines whether exposure is treated as fixed at baseline or time-updated. Because medication status can change during follow-up, time-fixed exposure definitions risk misclassification, whereas time-varying approaches reduce but do not eliminate bias. The three-year window provides a practical horizon for clinically relevant events, albeit one that may be insufficient for slower pathobiologic processes.

Covariate control typically includes demographic, metabolic, hepatic, and cardiovascular variables. Common adjustments include age, sex, diabetes, hypertension, dyslipidemia, body mass index, alcohol use patterns, baseline liver enzymes, noninvasive fibrosis scores, and comedications such as statins or other antithrombotics. Residual confounding remains a possibility even with careful modeling. Because aspirin is often indicated for atherosclerotic disease, differences in vascular risk at baseline can influence mortality unrelated to liver disease, while selective avoidance in patients with bleeding risk can shape hepatic event rates. The net observed associations thus reflect the interplay of indication, contraindication, and heterogeneity in underlying disease severity.

Reported associations

The central finding was an association between aspirin exposure alone and differences in risks of all-cause mortality and composite liver-related events across three years. The report did not present a treatment recommendation; rather, it characterized outcome patterns in a large, real-world MASLD population. Whether the observed differences translated into relative risk reductions or increases for specific endpoints is a matter for detailed review of the results section. Regardless of directionality, such signals inform hypotheses about the role of platelet inhibition or patient selection in shaping hepatic trajectories.

Outcome composites typically pool decompensation, variceal bleeding, hepatic encephalopathy, and liver-related hospitalization or mortality, while secondary analyses may consider incident hepatocellular carcinoma or transplant. The use of composites increases statistical power but may conflate mechanisms if components differ in pathogenesis or timing. For example, variceal bleeding may track portal pressure and hemostasis, whereas encephalopathy reflects ammonia handling and systemic inflammation. If associations differ across components, the composite estimate can obscure important nuance. Disaggregated results, when available, help clarify whether signals cluster in decompensation, cancer, or other domains.

All-cause mortality integrates hepatic and nonhepatic hazards, including cardiovascular events that are prevalent in MASLD. As such, mortality associations can reflect a mixture of aspirin effects on thrombosis risk, bleeding complications, infections, and progression of chronic liver disease. Competing risks are a key consideration: nonhepatic deaths may preclude observation of liver-related events, potentially altering incidence estimates if not modeled appropriately. Robust analyses often employ subdistribution hazards or cause-specific models to account for this dynamic, although approach details vary by dataset and objective.

Methodological considerations

Interpretation of observational associations rests on transparent exposure measurement and risk adjustment. Propensity methods, including propensity score matching or weighting, can balance measured confounders between exposed and unexposed groups. Nonetheless, unmeasured or poorly measured variables can persist, such as over-the-counter aspirin use not captured in health records, episodic alcohol intake, or dynamic disease activity. Modeling decisions about covariate selection and interaction terms can shift estimates, and sensitivity analyses help gauge robustness. Absent randomized allocation, cautious language about association rather than causation is warranted.

Immortal time and time-lag biases are frequent threats in medication-outcome analyses. If exposure is defined using future information, early person-time can be misclassified, producing a spurious protective association. Techniques that treat exposure as time-varying and anchor follow-up at well-defined baselines mitigate this concern. Awareness of immortal time bias is especially important when outcomes like death or first decompensation can occur early, compressing observation time in high-risk patients. Careful alignment of exposure windows, washout periods, and outcome ascertainment reduces such distortions.

Confounding by indication arises when the clinical reason for prescribing aspirin correlates with outcome risk. In MASLD, higher cardiovascular risk prompts aspirin use, yet that risk may also drive mortality independent of hepatic status. Conversely, clinicians may avoid aspirin in individuals with varices or thrombocytopenia, channeling lower bleeding risk into the exposed cohort and lowering observed liver-related events. Approaches to address confounding by indication include instrumental variable analyses, negative control outcomes, and sensitivity analyses for unmeasured confounding, though each has assumptions that must be examined.

Outcome misclassification is another consideration. Coding for decompensation, variceal bleeding, and encephalopathy may be specific but incomplete, and administrative capture can lag clinical reality. Validation with chart review or established algorithms improves confidence but can be resource intensive in multi-institutional settings. Mortality ascertainment varies by health system integration and linkage to national registries. If case definitions or capture probabilities differ across sites, pooled analyses must account for heterogeneity, including by random effects or site-level covariates.

Safety and bleeding risk

Any potential hepatic benefit of platelet inhibition must be weighed against the risk of gastrointestinal or portal hypertensive hemorrhage. Low-dose aspirin increases bleeding risk in proportion to mucosal vulnerability and the presence of varices, thrombocytopenia, or co-therapy with other antithrombotics. Proton pump inhibitors can mitigate upper gastrointestinal risk, though they do not address variceal bleeding. In advanced disease, endoscopic screening and management of varices inform background risk, and the balance of risks and benefits can be markedly different from earlier stages of steatosis or fibrosis. These clinical realities shape who receives aspirin in practice and therefore influence observed associations.

Beyond bleeding, considerations include acute kidney injury in the setting of cirrhosis and relative hypotension, as well as potential interactions with nonsteroidal anti-inflammatory drugs that can attenuate aspirin antiplatelet effects. Medication adherence fluctuates and is frequently undercaptured in electronic health records, leading to exposure misclassification. When exposure is intermittent, intention-to-treat and as-treated analyses may yield diverging estimates. Presenting both perspectives, when feasible, offers a more complete picture of the exposure effect pattern over time.

Mechanistic context

Platelet-hepatocyte and platelet-stellate cell interactions contribute to fibrogenic signaling in preclinical systems, and inhibition of thromboxane A2 synthesis may attenuate some profibrotic cascades. Aspirin also influences cyclooxygenase-2 and downstream lipid mediators that intersect with inflammatory networks central to MASLD pathogenesis. However, inflammation in MASLD is multifactorial, involving lipotoxicity, oxidative stress, and gut-liver axis signals. Any single pathway intervention is unlikely to dominate clinical outcomes across heterogeneous patient populations. Observational associations therefore serve as hypothesis generators that prioritize targets for randomized evaluation.

Hepatocellular carcinoma risk in MASLD emerges from a substrate of chronic inflammation and fibrosis, with additional contributions from metabolic factors such as insulin resistance. Antiplatelet strategies have been hypothesized to influence tumor surveillance and microenvironmental dynamics, though human evidence remains mixed. In parallel, aspirin can affect extrahepatic vascular outcomes that alter survival irrespective of liver disease. Disentangling these channels requires granular endpoint categorization and careful handling of competing risks, emphasizing why composite and cause-specific analyses should be interpreted together.

Clinical and research implications

For clinicians, the report offers updated real-world associations in an area where definitive trial data are limited. The key message is observational: aspirin exposure alone in MASLD correlated with differences in all-cause mortality and liver-related events over three years. Because direction and magnitude depend on the analytic specification and patient selection, translating these signals to individual decisions is premature. The findings will nonetheless inform shared discussions about risk, uncertainty, and monitoring when aspirin is used for cardiovascular indications in people at risk of hepatic complications.

For researchers, the analysis motivates targeted trials and advanced observational designs. Pragmatic randomized studies that enroll MASLD populations with stratification by fibrosis stage could clarify net clinical benefit or harm, including bleeding and decompensation. High-dimensional adjustment, emulation of target trials, and negative control outcomes can strengthen causal inference in real-world data when trials are impractical. Standardizing outcome definitions and leveraging adjudication frameworks improve comparability across datasets and institutions, building an evidence base that can support future guidelines.

Health systems and policymakers may focus on balancing medication safety with outcome monitoring in MASLD, where polypharmacy and multimorbidity are common. The intersection of hepatic and cardiovascular risks argues for integrated care pathways that include noninvasive fibrosis assessment and bleeding risk stratification in patients receiving long-term aspirin. Registries that capture longitudinal exposure, adherence, variceal status, and event details will enhance external validity and accelerate learning. As evidence matures, these infrastructures can support rapid-cycle evaluations that update risk estimates and inform practice.

In summary, a multi-institutional, three-year analysis reported an association between aspirin use alone and the risks of all-cause mortality and liver-related events in MASLD. The findings are timely and relevant but remain observational, shaped by exposure definitions, confounding, and competing risks. Safety considerations, especially bleeding, are integral to any interpretation. Next steps include rigorous causal inference in real-world data and pragmatic trials to determine whether and in whom aspirin modifies hepatic outcomes, while ensuring that clinical decisions remain anchored in established cardiovascular indications and individualized risk assessment.