O3018: ECG Abnormalities, Diastolic Hypotension, and Adverse Events After Systemic-pulmonary Artery Shunt

Background: A surgical systemic to pulmonary artery (SPA) shunt to provide pulmonary blood flow is a common technique for palliation of congenital heart disease. This shunt results in “runoff” from the systemic to pulmonary circulation during diastole. Because myocardial perfusion predominantly occurs during diastole, this can lead to myocardial ischemia, which may manifest as repolarization abnormalities on ECG. The threshold at which low diastolic blood pressure (DBP) after SPA shunt causes myocardial ischemia is unknown. We hypothesized that increasing grades of repolarization abnormality and increasing burden of diastolic hypotension in the early postoperative period would be associated with adverse outcomes.

Methods: This is a single-center, retrospective cohort study of patients with SPA shunt as their index surgical procedure from January 2010 to June 2020. ECGs were obtained within the first postoperative 72 hours (h) and graded based on repolarization abnormalities. Grade 2: ST elevation or depression 2 mm in 2 precordial leads or 1 mm in 2 limb leads; Grade 1: T wave inversion or flattening in 3 leads; Grade 0: did not meet criteria for grades 1 or 2. Invasive blood pressure monitoring was performed for the first postoperative 24h. Minimum and time-averaged DBPs were measured. We also specified three thresholds (25, 30, and 35 mmHg) and measured the time spent below each threshold in the first 24h. The primary outcome was a pre-discharge adverse event, defined as a composite of death, cardiac arrest, ECMO, reintervention to control pulmonary flow, and stage II or III necrotizing enterocolitis, that occurred after the first 24h postoperatively. The association of ECG grades and DBP with the outcome was evaluated with univariable logistic regression and multivariable models adjusting for chromosomal abnormalities, a known risk factor for adverse outcomes.

Results: 124 patients met inclusion criteria during the study period, and 20 (16.1%) had adverse outcomes. Of these, 109 had sufficient BP and ECG data for logistic regression analysis, with adverse outcomes occurring in 17 (15.6%). There were no significant differences in baseline characteristics between patients with and without adverse outcomes (Table 1). We observed no association between outcome and any measure of DBP or ECG abnormality on univariate analysis (Table 2). On multivariable regression, increasing time below 35 mmHg was associated with decreased odds of adverse outcome (OR 0.84, CI 0.69-0.98, p=0.04).

Conclusion: We observed no association between repolarization abnormalities on postoperative ECG and adverse outcomes, suggesting limitations of the one-time ECG as a marker of ischemia. We also found no association between adverse outcome and early diastolic hypotension below 25 or 30 mmHg and a decreased odds of adverse outcome with increasing time below 35 mmHg. This suggests that periods of diastolic hypotension may be well-tolerated, while elevated DBP may not always be protective.