Anticoagulants and myocardial infarction. The problems of pooling, drowning, and floating

Area of Science:

• Medical Research Methodology
• Clinical Epidemiology
• Biostatistics

Background:

• Retrospective data pooling from multiple studies is often used to address long-standing clinical questions due to the time and cost of randomized trials.
• A recent pooled analysis concluded anticoagulation benefits acute myocardial infarction (AMI) patients and questioned the ethics of further trials.
• However, the validity of this pooled analysis is questionable due to unmet scientific, statistical, and clinical standards.

Purpose of the Study:

• To critically evaluate the scientific, statistical, and clinical standards of a recent pooled analysis on anticoagulation in acute myocardial infarction.
• To highlight the potential pitfalls and limitations of retrospective data pooling when applied to clinical questions.

Main Methods:

• Review and critique of the methodology used in a specific pooled analysis of studies concerning anticoagulation in acute myocardial infarction.
• Assessment of the scientific validity criteria for data pooling, including patient similarity in diagnosis, severity, treatment, and outcomes.
• Evaluation of statistical pooling methods, emphasizing the use of exact numbers over percentages.
• Consideration of clinical applicability, particularly the impact of evolving ancillary therapies on the relevance of historical data.

Main Results:

• The pooled analysis failed to meet essential scientific criteria, as component studies lacked sufficient patient homogeneity.
• Statistical methods were inappropriate, relying on percentages instead of exact numbers for pooling.
• Clinical relevance was compromised by significant changes in ancillary therapies since the original studies were conducted.

Conclusions:

• The conclusions drawn from the "pooled" analysis regarding anticoagulation in acute myocardial infarction are disputed due to methodological flaws.
• Retrospective data pooling requires rigorous adherence to scientific, statistical, and clinical standards to yield reliable results.
• Investigators must exercise caution when pooling data, ensuring comparability across studies and considering the impact of contemporary medical practice.