Letter: Predictors and Outcomes of Pulmonary Embolism After Coronary Artery Bypass Grafting

To the Editor,

The recent analysis by Agrawal et al, ¹ which leverages the Nationwide Readmission Database (NRD) to assess predictors and clinical outcomes of pulmonary embolism (PE) after coronary artery bypass grafting (CABG), offers thought-provoking epidemiological findings. While this large-scale analysis provides valuable epidemiological insights, several methodological considerations warrant further discussion.

First, the study’s selection logic may limit the generalizability of its findings. By restricting the cohort to patients readmitted within 30 days, ¹ the analysis effectively truncates the observation of the full disease spectrum. As noted by Protopapas et al, ² post-CABG PE often exhibits significant early clustering. Furthermore, Bucerius et al ³ confirmed that most venous thromboembolic events occur within the first postoperative week. By excluding patients who developed PE during the index hospitalization or those who died before discharge, this readmission-based sampling may primarily reflect a “survivor” cohort rather than the true risk trajectory of the entire post-CABG population.

Second, administrative databases lack the clinical granularity necessary to capture perioperative determinants. The NRD omits critical physiological parameters that influence coagulation homeostasis, such as cardiopulmonary bypass (CPB) duration, which is a recognized trigger for systemic inflammation and thrombosis, ⁴ and baseline cardiac function (e.g., LVEF). Most critically, the study lacks detailed records of perioperative anticoagulation prophylaxis. As emphasized by Montrief et al, ⁵ variations in management pathways significantly alter outcomes. Without adjusting for these direct interventional variables, identified predictors may reflect differences in preventive medical practices rather than inherent biological susceptibility.

Third, significant heterogeneity in baseline characteristics poses an interpretative challenge. Agrawal et al ¹ reported that the PE group carried a lower comorbidity load, while the non-PE group more frequently had sepsis and acute kidney injury. This “healthy bias” suggests the control group consisted of more critically ill patients. In clinical statistics, an excessively high baseline risk in the reference group can dilute the target factor’s (PE) true contribution to mortality. Consequently, the reported 1.45-fold mortality risk may be an underestimate of PE’s actual harm.

Finally, clinical recommendations must be more circumspect. The authors suggest aggressive anticoagulation starting on postoperative day 1. However, Sandner et al ⁶ emphasized that anticoagulation timing must be individualized to balance graft patency against bleeding risks. According to Singh et al, ⁷ the early postoperative period is a sensitive window of trade-offs between thrombus prevention and severe hemorrhage (e.g., fatal tamponade). Without data on bleeding events, proposing universal early aggressive anticoagulation requires more rigorous evidentiary support.

In conclusion, future research utilizing prospective registries with higher granularity is essential to balance the dual risks of thrombosis and hemorrhage in optimizing preventive strategies.