Clinical Scorecard: Synchronizing Time Zero and Treatment Assignment in Target Trial Emulation of Left Ventricular Unloading

At a Glance

Key Highlights

• Eligibility, treatment assignment, and follow-up should be aligned at time zero; a 48-hour ECMO-duration criterion may select patients based on post-baseline survival and support.
• Classifying unloading by receipt within 24 hours while starting follow-up at cannulation creates a grace-period concern because patients must survive until device placement.
• Cloning, artificial censoring, and inverse-probability-of-censoring weighting are described as methods to align strategies during a treatment grace period.
• Unloading decisions were driven by evolving clinical findings that may predict outcomes and be affected by ECMO flow and inotropic therapy.
• The weighted risk difference was 1.4% (95% CI, −11.8% to 14.6%); the authors state this indicates no detectable association in the weighted overlap population, not equivalence.

Guideline-Based Recommendations

Diagnosis

• The letter reports pulmonary edema, left ventricular dilation, elevated pulmonary artery occlusion pressure, very low aortic velocity-time integral, and continued failure of aortic valve opening as indications for unloading.

Management

• The letter describes the EARLY-UNLOAD trial as comparing routine early unloading with a conventional approach permitting rescue unloading when prespecified signs of increased left ventricular afterload emerged.
• The authors propose emulating clinically defined dynamic strategies using time-updated hemodynamic and echocardiographic measurements.

Monitoring & Follow-up

• The authors identify time-updated hemodynamic and echocardiographic findings as relevant to the evolving unloading decision.
• Aortic velocity-time integral data were unavailable from one center.

Risks

• If applied after cannulation, a 48-hour ECMO-duration eligibility criterion may omit patients who die or are weaned early and create post-baseline selection bias.
• A 24-hour exposure window with follow-up beginning at cannulation may create a grace-period survival bias.
• The authors state that baseline propensity-score weighting cannot address time-varying confounding or near-deterministic coupling of device and center.

Patient & Prescribing Data

Patients with cardiogenic shock receiving VA-ECMO; the article discusses a weighted overlap population receiving local treatment practices.

The analysis compared left ventricular unloading with VA-ECMO alone. IABP use occurred almost exclusively in Paris and Impella use largely in Hamburg, limiting separation of device, center, and treatment-policy effects.

Clinical Best Practices

• The authors argue that eligibility, treatment assignment, and follow-up initiation should be temporally aligned in target trial emulation.
• Clarify whether a 48-hour ECMO-duration criterion was used as an eligibility condition.
• The authors describe cloning, artificial censoring, and inverse-probability-of-censoring weighting as methods to align strategies during a treatment grace period.
• The authors note that a 24-hour landmark analysis would estimate an effect only among patients alive and receiving support at that landmark.
• The authors recommend accounting for time-varying clinical triggers and center-level treatment practices when comparing unloading strategies.

Related Resources & Content