Quantifying high-pass filter-induced ST-segment distortion

Standard 12‑lead electrocardiograms (ECGs) use a 0.05 Hz high-pass filter, while continuous bedside monitoring systems commonly employ higher filter settings (≥0.5 Hz) to reduce baseline wander. Building upon foundational studies of filter-induced artifact, we aimed to apply modern parametric modeling to quantify the magnitude, anatomical distribution, and clinical determinants of this phenomenon.

We conducted a repeated-measures analysis of 150 patients referred for electrophysiology study. Standard 12‑lead ECGs were analyzed using four high-pass filter settings: 0.05 Hz, 0.1 Hz, 0.5 Hz, and 1.0 Hz. Linear mixed-effects models assessed the effects of filter frequency and baseline QRS morphology on ST-segment deviation. Logistic regression was used to determine the odds of clinically significant false-positive ST elevation (≥1.0 mm).

Increasing high-pass filter frequency produced significant, frequency-dependent ST-segment elevation, with a pronounced "ceiling effect" occurring at the standard telemetry default of 0.5 Hz. The artifact was anatomically selective, predominantly affecting the right precordial leads (maximum increase at 1.0 Hz: V1 + 0.51 mm, V2 + 0.80 mm; both p < 0.001), while limb leads were minimally affected (Lead I + 0.09 mm, Lead II +0.19 mm). An R/S ratio < 1 in the right precordial leads significantly amplified artifact magnitude at 1.0 Hz. Use of a 1.0 Hz filter independently increased the odds of false-positive ST elevation ≥1.0 mm nearly ten-fold in V1 (OR: 9.14, 95% CI: 4.16-20.11; p < 0.0001) and five-fold in V2 (OR 4.96, 95% CI 2.87-8.58; p < 0.001).

High-pass filter settings ≥0.5 Hz induce substantial, anatomically selective ST-segment elevation, highlighting the need for caution when interpreting ST segments on bedside monitoring systems.