Simple diagnosis of limb-lead reversals by predictable changes in QRS axis

Background: Limb-lead reversals (LLRs) remain clinically problematic. Because the frontal QRS axis is derived from an equilateral Einthoven triangle and LLRs either rotate (180 degrees horizontally (mirror-image (M)) and/or 120 degrees vertically (clockwise (C)/counterclockwise (CC)) or distort the triangle (by forcing a bipolar lead to record across the lower extremities (LE) where electrical potentials approach zero (zero-potential lead)), we hypothesize that LLR axes changes from a baseline value (n) are predictable. Methods: Three hundred and sixty ECGs with all 24 limb-lead combinations from 15 individuals were analyzed. Predicted and actual QRS axes were compared using linear regression. Results: Twenty-four lead combinations produced only 12 (11 abnormal) ECG patterns and diagnosis depended upon identifying upper extremity (UE) cable configurations. Predicted formulas for rotation-type LLR axes (M, C, CC, MC, and MCC) were 180 - n, n - 120, n + 120, 300 - n, and 60 - n, respectively. Corresponding mean differences between predicted and actual values were 4 +/- 5 degrees, 4 +/- 2 degrees, 7 +/- 10 degrees, 5 +/- 7 degrees, and 3 +/- 4 degrees (all r = 0.99 - 1.00, P < 0.0001). The predicted formula for distortion-type LLR axes is (zero-potential lead axis) +/- 90 degrees. Actual mean values for zero-potential lead I, II, and III were 90 +/- 7 degrees or 266 +/- 17 degrees, -31 +/- 6 degrees or 148 +/- 14 degrees, and 26 +/- 14 degrees or 209 +/- 9 degrees, respectively. The mean difference between predicted and actual values for all LLRs was 5 +/- 8 degrees (r = 1.00, P < 0.0001). Conclusions: LLR axes are predictable within an average of 5 degrees. This might help differentiate an acute axis shift due to an LLR from serious medical conditions that may require treatment.