CRT-D at 148 Milliseconds: A QRS Duration Two Ticks Below the Guideline's Sharpest Line
Left bundle branch block, EF 30%, and a QRS duration reading two milliseconds short of the guideline's Class I threshold for cardiac resynchronization. His GDMT is also still mid-titration — two real reasons the decision isn't as simple as reading the number off the ECG.
L.M., a 70-year-old man, drove a city bus for thirty-four years before retiring, and still keeps touch with several of his regular riders from the route he drove longest. He was diagnosed with nonischemic HFrEF four months ago, EF 30%, and has been on quadruple GDMT since — sacubitril/valsartan (currently at half of target dose, still being uptitrated), a beta-blocker near target, spironolactone at target, and an SGLT2 inhibitor at target. Despite that, he remains NYHA class III: stairs leave him genuinely short of breath, and he's had to give up the twice-weekly walks with his old riders that used to be the best part of his week. His ECG shows left bundle branch block with a QRS duration of 148 milliseconds — two milliseconds short of the 150ms threshold that would place him in the guideline's strongest recommendation category for cardiac resynchronization.
The ACC/AHA/HRS criteria for CRT-D draw a real distinction at that threshold: LBBB with QRS 150ms or greater is a Class I recommendation, "should be done," in patients with his EF and symptom profile on GDMT; LBBB with QRS in the 120-149ms range drops to Class IIa, "can be useful" — a real, positive recommendation, just a step down in strength. LBBB morphology itself is the single strongest predictor of CRT response across the trial evidence establishing these criteria, more consistently than QRS duration alone. Two separate complications sit alongside the numerical question: his GDMT isn't yet fully optimized, since ARNI uptitration is still in progress and both EF and QRS duration can shift with further medical therapy; and QRS duration itself carries real measurement variability between machines and readers, on the order of a few milliseconds — meaningful when the guideline's line falls this close to his own number.
Heart failure/electrophysiology joint clinic
I'd proceed with CRT-D now. LBBB is the strongest single predictor of response across the trial evidence — stronger than QRS duration alone — and the Class IIa recommendation at 148ms is a real, positive recommendation, not a "wait and see." He's two milliseconds from Class I, not on the other side of a meaningful biological line.
I'm not saying the class distinction is meaningless — I'm saying it shouldn't be read as "will respond" versus "won't."
His GDMT isn't finished yet — the ARNI is still at half target dose, and both EF and QRS duration can genuinely shift with further optimization. The guideline's own expectation is a period of stable, optimized therapy before finalizing device candidacy. I'd rather complete titration first than implant a device a few more months of medical therapy might change the calculus on.
I'm not disputing the trial evidence on LBBB — I'm saying his own numbers aren't settled yet, and that matters independent of which recommendation class he falls into today.
Before either of you finalizes a position off today's number, I'd want a manual re-measurement of that QRS duration, not just the automated reading. Machine and reader variability on the order of a few milliseconds is real, and a boundary this sharp deserves a number this examined before it drives anything.
Agreed: complete ARNI uptitration to target over the next 4-6 weeks, obtain a manual QRS re-measurement today, and reassess CRT-D candidacy at that point with both the medical therapy and the measurement question settled.
Not agreed, and carried forward explicitly rather than smoothed over:
All three voices agreed this becomes a Class I recommendation once GDMT is also complete — the Electrophysiologist's original instinct to proceed sooner would look more clearly right in hindsight.
The Heart Failure Cardiologist's caution is reinforced — CRT-D candidacy would need to be reassessed against his new baseline, not his numbers from today.
The Electrophysiologist's LBBB-predicts-response point was accepted by both other voices as real and not in dispute — the delay agreed to today is about completing GDMT and confirming the measurement, not about doubting that he would likely respond to CRT-D if implanted now.