Cutting off half of QRS duration can cause overcorrection of QT interval in left bundle branch block

Dear Editor,

We read with great interest the article by Bogossian et al., (2017) entitled “A New Experimentally Validated Formula to Calculate the QT Interval in the Presence of Left Bundle Branch Block Holds True in the Clinical Setting”. March, 2017. The authors demonstrated that in the presence of left bundle branch block (LBBB), the true QT interval could be estimated using the formula [QTm = QTb − 50% × QRSb] (Bogossian et al., 2014). We tested this formula and found it actually overcorrected QT interval in patients with LBBB.

The falsely prolonged QT interval in LBBB is primarily caused by the widened QRS (QRSb) due to delayed depolarization (Bogossian et al., 2014; Das, 1990). The key is to know how much of QRSb added to the QT interval in LBBB. In the formula, the authors simply chopped off 50% of QRSb. Table 1 demonstrates that when we applied this formula to a group of patients (n = 22) with intermittent LBBB, the estimated QT becomes significantly shorter than the baseline level (424 ± 31 vs. 441 ± 34 ms, p = .004). The JT interval did not alter much in LBBB, so what is the cause? We looked at the net increase in QRS duration, and noticed it was only about 40% of QRSb (0.39 ± 0.04). The same is true in the authors’ data (Bogossian et al., 2017). Therefore, we modified the formula by cutting off 40% of QRSb, and by doing so the best estimate of the true QT interval was achieved (442 ± 33 vs. 441 ± 34 ms, p = .969). In other words, replacing 50% with 40% of QRSb in the formula [QTm = QTb − 40% × QRSb] can prevent overcorrection of the QT interval in LBBB. To determine whether such modification is valid, we applied Rautaharju formula designed for LBBB (Rautaharju, Zhang, Prineas, & Heiss, 2004) in our dataset. Table 1 shows that the results are compatible. Nevertheless, Rautaharju formula is more complicated. QTm = QTb − 40% × QRSb, on the other hand, is much easier to remember and apply in clinical practice.

Table 1.

Estimating QT interval in patients with intermittent LBBB

Case	Narrow QRS						Left bundle branch block
	HR, bpm	QRS, ms	JT, ms	JTc, ms	QT, ms	QTc, ms	HR, bpm	QRS, ms	JT, ms	JTc, ms	QT, ms	QTc, ms	ΔQRS/QRSb	QTc50, msa	QTc40, msb	QTcR, msc
1	107	90	255	371	345	461	83	160	295	374	455	534	0.44	440	459	456
2	91	95	305	397	400	492	98	160	295	423	455	583	0.41	480	501	474
3	74	90	305	348	395	438	60	150	343	343	493	493	0.40	418	433	448
4	74	85	315	358	400	443	76	140	300	357	440	497	0.39	418	433	438
5	83	100	260	323	360	423	97	160	240	348	400	508	0.38	406	427	417
6	69	78	302	331	380	409	83	130	270	341	400	471	0.40	395	410	418
7	68	105	375	407	480	512	69	165	340	376	505	541	0.36	453	471	479
8	71	95	305	341	400	436	73	155	295	340	450	495	0.39	410	427	440
9	48	95	395	345	490	440	52	160	375	339	535	499	0.41	424	439	470
10	79	80	280	333	360	413	81	140	270	335	410	475	0.43	394	410	427
11	67	80	310	331	390	411	68	150	320	350	470	500	0.47	420	436	444
12	94	105	275	370	380	475	97	175	265	384	440	559	0.40	448	470	443
13	79	80	300	356	380	436	106	135	245	371	380	506	0.41	416	434	429
14	48	100	310	268	410	368	48	180	340	286	520	466	0.44	385	402	422
15	67	95	325	348	420	443	68	140	320	349	460	489	0.32	415	429	455
16	85	85	315	390	400	475	136	120	200	362	320	482	0.29	392	410	402
17	72	95	345	387	440	482	88	170	345	455	515	625	0.44	521	542	514
18	71	80	300	335	380	415	63	125	330	343	455	468	0.36	404	417	443
19	54	95	335	311	430	406	66	160	320	342	480	502	0.41	418	435	440
20	81	80	280	338	360	418	74	140	295	343	435	483	0.43	406	421	442
21	65	110	340	359	450	469	73	160	325	376	485	536	0.31	447	465	459
22	79	100	290	347	390	447	80	160	290	360	450	520	0.38	427	446	435
Mean	74	92	310	350	402	441	79	152	301	359	452	510	0.39	424d	442e	445f
SD	14	10	34	31	38	34	20	16	41	33	50	39	0.04	31	33	24

^aQTc₅₀: QTm = QTb − 50% × QRSb was applied for the QT component of heart rate‐corrected QT interval by cutting off 50% of QRS duration in LBBB.

^bQTc₄₀: QTm = QTb − 40% × QRSb was applied for QTc, respectively.

^cQTc_R: Heart rate−corrected QT interval corrected by Rautaharju formula. (Rautaharju et al., 2004).

^dPaired t test between QTc₅₀ and baseline QTc with narrow QRS: p = .004.

^ePaired t test between QTc₄₀ and QTc with narrow QRS: p = .969.

^fPaired t test between QTc₅₀ and QTc with narrow QRS: p = .533.

Based on our testing, the net increase in QRS is about 40% of the widened QRS in LBBB in both our and the authors’ dataset. Therefore, we would suggest adjusting the quotient from 50% to 40% of QRSb in the new formula to prevent overcorrection of QT interval.