Utility of Provocative Testing in the Diagnosis and Genotyping of Congenital Long QT Syndrome: A Systematic Review and Meta‐Analysis

Abstract

Background

Diagnosis is particularly challenging in concealed or asymptomatic long QT syndrome (LQTS). Provocative testing, unmasking the characterization of LQTS, is a promising alternative method for the diagnosis of LQTS, but without uniform standards.

Methods and Results

A comprehensive search was conducted in PubMed, Embase, and the Cochrane Library through October 14, 2021. The fixed effects model was used to assess the effect of the provocative testing on QTc interval. A total of 22 studies with 1137 patients with LQTS were included. At baseline, QTc interval was 40 ms longer in patients with LQTS than in controls (mean difference [MD], 40.54 [95% CI, 37.43–43.65]; P<0.001). Compared with the control group, patients with LQTS had 28 ms longer ΔQTc upon standing (MD, 28.82 [95% CI, 23.05–34.58]; P<0.001), nearly 30 ms longer both at peak exercise (MD, 27.31 [95% CI, 21.51–33.11]; P<0.001) and recovery 4 to 5 minutes (MD, 29.85 [95% CI, 24.36–35.35]; P<0.001). With epinephrine infusion, QTc interval was prolonged both in controls and patients with QTS, most obviously in LQT1 (MD, 68.26 [95% CI, 58.91–77.60]; P<0.001) and LQT2 (MD, 60.17 [95% CI, 50.18–70.16]; P<0.001). Subgroup analysis showed QTc interval response to abrupt stand testing and exercise testing varied between LQT1, LQT2, and LQT3, named Type Ⅰ, Type Ⅱ, and Type Ⅲ.

Conclusions

QTc trend Type Ⅰ and Type Ⅲ during abrupt stand testing and exercise testing can be used to propose a prospective evaluation of LQT1 and LQT3, respectively. Type Ⅱ QTc trend combined epinephrine infusion testing could distinguish LQT2 from control. A preliminary diagnostic workflow was proposed but deserves further evaluation.

Nonstandard Abbreviations and Acronyms

CEs

cardiac events

LQTS

long QT syndrome

MD

mean difference

Clinical Perspective

What Is New?

• Provocative testing, unmasking the characterization of long QT syndrome (LQTS), is a promising alternative method for the diagnosis of concealed or asymptomatic LQTS, but without uniform standards.

• Our study is the first attempt to determine the pooled utility of provocative testing in the diagnosis and genotyping of LQTS, and a preliminary diagnostic workflow was proposed, but deserves further evaluation.

What Are the Clinical Implications?

• A preliminary diagnostic workflow of LQTS may be helpful for the diagnosis and genotyping of LQTS, which is very important for the treatment and prognostic estimation of LQTS.

Congenital long QT syndrome (LQTS) is a potentially lethal cardiac channelopathy, characterized by prolonged QT interval on electrocardiography and predisposition to ventricular arrhythmia and sudden cardiac death. ¹ To date, >17 genes have been identified as LQTS‐causing genes, but recent expert opinion disputes the validity of several of them. The most common subtypes are caused by mutations in KCNQ1 (LQT1), KCNH2 (LQT2), and SCN5A (LQT3), accounting for ≈95% of genotype‐positive LQTS cases, whereas 75% to 80% account for clinically definite LQTS cases. ² , ³ Schwartz score, combining characterizations of electrocardiography abnormalities, clinical parameters, and family history, is widely used in the diagnosis of LQTS. ⁴ However, the diagnosis is particularly challenging in patients who were asymptomatic (asymptomatic LQTS) or had normal to borderline QTc interval at rest (concealed LQTS), but were still at significant risk of ventricular arrhythmia and sudden cardiac death. ⁵ Consequently, there is an urgent need for an alternative method for the diagnosis of LQTS.

Provocative testing, including exercise stress testing (treadmill test, ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ bicycle test ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ ), abrupt stand testing, ⁵ , ⁶ , ⁸ , ¹⁹ , ²⁰ , ²¹ , ²² and epinephrine QT stress testing, ²³ , ²⁴ , ²⁵ , ²⁶ has been explored as a means of unmasking the characterization of patients with LQTS. QTc interval changes in provocative testing have been proposed as meaningful parameters to facilitate diagnosis and genotyping of LQTS but without uniform standards between studies. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ Therefore, we performed a systematic review and meta‐analysis of currently available medical literature, for 3 objectives: (1) to validate the differences between patients with LQTS and healthy population in provocative testing; (2) to explore the characteristics of patients with LQT1, LQT2, and LQT3 in provocative testing; and (3) to propose a diagnostic workflow for diagnosis and genotyping of LQTS.

METHODS

Search Criteria

All authors declare that the supporting data are available upon reasonable request. Our systematic review and meta‐analysis were performed in accordance with the guidelines from the Preferred Reporting Items of Systematic Review and Meta‐Analysis ²⁷ and Meta‐analysis of Observational Studies in Epidemiology. ²⁸ A comprehensive search was conducted by using the electronic database PubMed, Embase, and the Cochrane Library without language restriction through October 14, 2021. The electronic search terms were as follows: provocative test, exercise, treadmill, bicycle, stand, epinephrine, and long QT syndrome. The reference lists of trials, systematic reviews, and meta‐analyses were also manually screened for additional studies that fit our inclusion criteria. All citations were initially identified at the level of title and abstract, and the candidate list of articles was confirmed by reviewing the full text. Two authors (Y.Y. and L.T.T.) checked the citations separately, and had a discussion with the corresponding author (Z.P.) to achieve consensus if controversies existed.

Inclusion Criteria and Exclusion Criteria

Studies were included in this meta‐analysis if (1) patients with LQTS and control (healthy population or low probability of LQTS ⁵ , ¹³ , ¹⁶ , ²⁴ ) were both included in the study; (2) LQTS was diagnosed by clinical features and/or cardiac channel gene screening; (3) provocative testing was performed in both patients with LQTS and control; and (4) provocative testing included abrupt stand testing, treadmill test, bicycle test, or epinephrine infusion testing. Studies were excluded if (1) <10 patients with LQTS enrolled; (2) studies included either patients with LQTS or control; (3) the most appropriate studies were included, ⁷ , ⁹ , ¹³ , ¹⁶ , ¹⁹ , ²⁴ , ²⁵ , ²⁶ then duplicate groups were excluded ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ ; (4) studies in which QTc interval or ΔQTc could not be extracted, and conference abstracts were excluded; and (5) provocative testing including adenosine exposure, ³⁸ mental stress, ¹⁸ , ³⁹ Valsalva maneuver, ⁴⁰ , ⁴¹ and cold pressor test ⁷ , ⁴¹ , ⁴² were excluded because fewer studies were insufficient for analysis.

Data Extraction and Quality Assessment

We used a predefined standard data‐extraction form to collect information by 1 author (Y.Y.) and verified by another author (L.T.T.). The information for each trial included author, study design, enrolled population, sex, age, resting QTc, QTc formula, subtypes of LQTS, patients LQTS with cardiac events (CEs) and β‐blocker, and types of provocative testing and specific protocols. QTc intervals at baseline, abrupt stand, peak exercise and recovery 4 to 5 minutes were extracted; recovery 4 to 5 minutes represented late recovery in exercise testing. Data in the scatter plots were extracted by GetData Graph Digitizer software. Of the included studies, 4 analyzed overt or concealed patients with LQTS separately, ⁸ , ¹⁸ , ²³ , ²⁶ and the rest enrolled both overt and concealed LQTS. ⁵ , ⁶ , ⁷ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁹ , ²⁴ , ²⁵ To explore the pooled effect of QTc at different time points during provocative testing, ΔQTc was used to balance the differences of baseline QTc interval. ΔQTc was calculated by QTc interval at any time point minus QTc interval at supine before the test. The NIH Quality Assessment Tool for Case Series Studies was used to assess study quality. ⁴³

Statistical Analysis

Review Manager 5.3 (Cochrane Collaboration, Oxford, UK) was used for all data analyses. We used the fixed‐effects model to pool data, assuming that the underlying effects in contributing studies are identical. ⁴⁴ Differences of QTc or ΔQTc between groups were expressed as mean difference (MD) with a 95% CI. Heterogeneity across each meta‐analysis was evaluated by I², and values >50% were considered as a high degree of between‐study statistical heterogeneity. P<0.05 was considered statistically significant.

RESULTS

Study Characteristics

Two authors independently screened titles and abstracts of 2253 citations yielded from electronic search. After removing the duplicate and irrelevant articles, 22 studies met the predefined inclusion criteria and were included in our meta‐analysis (Figure 1A). Finally, a total of 1137 patients with LQTS were included; 42% were asymptomatic LQTS, 18% were symptomatic, and clinical symptoms were not mentioned in 40% of patients (Figure 1B). In terms of LQTS subtypes, the percentage of LQT1, LQT2, and LQT3 were 50%, 39%, and 7%, respectively (Figure 1C). The characteristics of all included studies are summarized in Table 1. Among these studies, 9 studies evaluated the effect of treadmill exercise testing, ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ 3 of which completed abrupt stand testing at the same time, ⁵ , ⁶ , ⁸ 5 focused on bicycle test, ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ 4 performed abrupt stand testing, ¹⁹ , ²⁰ , ²¹ , ²² and 4 concerned epinephrine QT stress testing. ²³ , ²⁴ , ²⁵ , ²⁶ Of enrolled patients with LQTS, the age ranged from 0 to 64 years old, ≈20% with CEs, and >21% with β‐blocker treatment. Details of another provocative testing (adenosine exposure, ³⁸ mental stress, ¹⁸ , ³⁹ Valsalva maneuver, ⁴⁰ , ⁴¹ and cold pressor test ⁷ , ⁴¹ , ⁴² ), which is excluded from our meta‐analysis, have been presented in Table 2.

Figure 1. Flowchart of literature search and distribution of enrolled LQTS population.

A, Flowchart of literature search; (B) Distribution of symptoms of patients with LQTS; (C) Distribution of subtypes of LQTS, others included other subtypes, unclassified and LQTS subtype not applicable. LQTS indicates long QT syndrome; and NA, not applicable.

Table 1.

Characteristics and Quality Analysis of Included Studies

Author, y	Design	Population, n	Male, n	Age, y	QTc formula	Patients with LQTS (n)				Testing	Protocols	Quality assessment
						Resting QTc (ms)	Subtype	CEs	β‐blocker
Roston,2021 6	Re	Asymptomatic LQT 14 Control 9	7 1	<19	Bazett’s	475 (434–500)	LQT1: 9 LQT2: 4 LQT8: 1	0	0	Treadmill exercise test QT‐stand test	BCCH or Bruce	Good
Takahashi, 2020 7	Re	Asymptomatic LQT 20 Control 22	9 8	11.2±3.4	Fridericia’s	510±43	LQT1: 3 LQT2: 3 LQT3: 12 LQT7: 1 LQT8: 1	0	0	Treadmill exercise test Cold‐water face immersion	Bruce	Good
Patel, 2020 8	Re	Concealed LQT 21 Control 76	9 44	≤21	Bazett’s	446±23	LQT1: 13 LQT2: 5 LQT3: 3	Syncope: 5 VA: 2	15	Treadmill exercise test	Bruce	Good
Wu 2013 9	Re	LQT 18 Control 45	5 19	8–47 8–41	Bazett’s	521±65	LQT1: 3 LQT2: 15	NA	NA	Treadmill exercise test	Bruce	Fair
Sy, 2011 5	Pro	Asymptomatic LQT 48 Control 21	26	35±18	Bazett’s	477±38	LQT1: 28 LQT2: 20	0	23	Treadmill exercise test or bicycle ergometry test	Bruce NA	Good
Horner, 2011 10	Re	LQT 155 Control 88	59 38	25±14 21±13	Bazett’s	468±31	LQT1: 82 LQT2: 55 LQT3: 18	NA	71	Treadmill exercise test	Bruce	Good
Wong, 2010 11	NA	LQT 95 Control 64	33 25	Median 26	Bazett’s	LQT1 461 (43) LQT2 482 (39)	LQT1: 50 LQT2: 45	NA	NA	Treadmill exercise test, bicycle ergometry test	Bruce Burst and gradual	Good
Takenaka, 2003 12	NA	LQT 82 Control 33	27 17	Mean 28–32	Bazett’s	LQT1 510±68 LQT2 520±61	LQT1: 51 LQT2: 31	47	0	Treadmill exercise test	Bruce	Good
Shimizu,1991 13	Re	LQT 11 Control 12	3 5	10–45	Bazett’s	555±60	NA	10	NA	Treadmill exercise test Isoproterenol infusion Atrial pacing	Bruce	Fair
Vink, 2021 20	Pro	LQT 47 Control 86	18 47	12 (8–15) 10 (7–14)	Bazett’s	466±36	LQT1: 26 LQT2: 19 LQT3: 2	3	9	Abrupt standing test	Viskin protocol	Good
Muñoz‐Esparza, 2017 21	Pro	LQT 36 Control 41	18 19	36±17 39±15	Bazett’s	450±31	LQT1: 6 LQT2: 20 LQT7: 3 Unclassified: 7	N/A	26	Abrupt standing test	Viskin protocol	Fair
Ben Bassat, 2013 22	pro	LQT 69 Control 82	18 46	31±14 35±10	Bazett’s	471±39	LQT1: 31 LQT2: 28 LQT7: 3 Unclassified: 6	N/A	N/A	Abrupt standing test	Viskin protocol	Good
Adler,2012 19	Re	LQT 108 Control 112	41 59	30±16 33±12	Bazett’s	469±40	LQT1: 46 LQT2: 47 LQT3: 12 Unclassified: 3	28	NA	Abrupt standing test	Viskin protocol	Good
Charisopoulou,2019 14	Re	LQT 47 Control 35	22 16	45±15 47±13	Bazett’s	LQT 453±42	LQT1: 36 LQT2: 11	20	25	Bicycle ergometer test	Gradual increase	Good
Aziz, 2011 15	NA	LQT 50 Control 108	20 45	≤21	Bazett’s	LQT1 452±33 LQT2 475±35	LQT1: 29 LQT2: 21	13	28	Bicycle ergometer test	Gradual increase	Good
Chattha, 2010 16	NA	LQT 50 Control 25	18 9	33±16	Bazett’s	LQT1: 467±32 LQT2: 465±25	LQT1: 25 LQT2: 25	14	28	Bicycle ergometry test	Burst and gradual increase	Good
Walker, 2005 17	NA	LQT 31 Control 31	8 9	16–45	Bazett’s	Concealed: 508±36 Overt: 496±69	LQT1: 3 LQT2: 13 Unknown: 15	17	24	Bicycle ergometer test	Burst	Good
Paavonen, 2001 18	Re	Asymptomatic LQT 30 Control 14	13 8	13–59 22–58	Fridericia’s	LQT1 429±31 LQT2 417±43	LQT1: 16 LQT2: 14	0	0	Bicycle ergometer test Mental stress	Burst and gradual increase	Good
Hekkala, 2011 23	NA	Asymptomatic LQT 30 Control 15	15 7	20–53	Bazett’s	LQT 413±25	LQT1: 10 LQT2: 10 LQT3: 10	0	0	Epinephrine Test	Bolus infusion	Good
Vyas, 2006 24	Re	LQT 81 Control 44	34 14	8–59	Bazett’s	LQT1 456 (397–517) LQT2 486 (388–644) LQT3 473 (424–532)	LQT1: 40 LQT2: 30 LQT3: 11	NA	0	Epinephrine test	Incremental infusion	Good
Shimizu, 2004 25	Pro	LQT 60 Control 30	24 14	4–64	Bazett’s	LQT1 470±41 LQT2 503±33 LQT3 506±41	LQT1: 31 LQT2: 23 LQT3: 6	28	0	Epinephrine test	Bolus infusion+continuous	Good
Shimizu, 2003 26	NA	Overt LQT 19 Concealed LQT 15 Control 15	5 8 6	Mean 22–31	Bazett’s	Overt 507±31 Concealed 427±21	LQT1: 34	15	0	Epinephrine test	Bolus+continuous infusion	Good

BCCH, British Columbia Children's Hospita; CEs, cardiac events (syncope, sudden cardiac arrest, or ventricular arrhythmia); HR, heart rate; LQTS, long QT syndrome; NA, not applicable; Pro, prospective; and Re, retrospective.

Table 2.

Characteristics of Studies With Other Provocative Tests

Author, y	Population, n	Age, y	LQTS subtype (n)	Testing	QTc in control (ms)		QTc in LQTS (ms)
					Before	After	Before	After
Viskin, 2006 38	LQT 18 Control 20	LQT 32.7±15.8 Control 31.4±12.2	NA	Adenosine test	399±29	458±58	454±24	569±53
Etienne, 2018 39	LQT 36 Control 34	Median 41±2.1	LQT1: 20 LQT2: 16	Mental stress test	ΔQTc 24 (10, 40)		ΔQTc 67 (44, 91)
Paavonen, 2001 18	LQT 30 Control 14	LQT1 35±13 LQT2 41±1 Control 43±10	LQT1: 16 LQT2: 14	Mental stress test	ΔQTf 24±26		LQT1 ΔQTf 27±36 LQT2 ΔQTf 30±47
Mitsutake, 1981 40	LQT 8 Control 9	LQT 39±7 Control 27±2	NA	Valsalva strain	500±30	570±60	450±20	480±20
Eggeling, 1993 41	LQT 14 Control 14	NA	NA	Valsalva maneuver	412±25	407±18	489±55	497±49
				Cold pressor test		407±19		512±45
Takahashi, 2020 7	LQT 20 Control 22	Mean 11.2±3.4	LQT3: 12 Non‐LQT3: 8	Cold‐water face immersion	/	Ratio of min/max HR 0.47 (0.43, 0.52)	/	Ratio of min/max HR 0.49 (0.45, 0.51)
Yoshinaga, 1999 42	LQT 10 Control 48	LQT 11.7±2.8 Control 12.6±2.3	NA	Cold‐water face immersion	420±20	390±30	560±60	600±30
Shimizu, 1991 13	LQT 11 Control 12	LQT 10–45 Control 15–28	NA	Isoproterenol infusion	399±10	436±13	466±50	556±33
				Atrial pacing	400±21	426±18	476±57	486±59

HR, heart rate; LQTS, long QT syndrome; and NA, not applicable.

Effect of Exercise Stress Testing and Abrupt Stand Testing on QTc Interval

To determine the pooled effect of exercise stress testing and abrupt stand testing on QTc interval, we performed forest plots at baseline, peak exercise, recovery 4 to 5 minutes, and abrupt stand. As shown in Figure 2A, at baseline, the QTc interval was 40 ms longer in patients with LQTS than in control (MD, 40.54 [95% CI, 37.43–43.65]; I²=94%, P<0.001). Compared with control group, patients with LQTS had nearly 30 ms longer ΔQTc both at peak exercise (MD, 27.31 [95% CI, 21.51–33.11]; I²=89%, P<0.001) and recovery 4 to 5 minutes (MD, 29.85 [95% CI, 24.36–35.35]; I²=92%, P<0.001), as presented in Figure 2B and 2C. As shown in Figure 2D, patients with LQTS had ≈29 ms longer ΔQTc at the abrupt stand (MD, 28.82 [95% CI, 23.05–34.58]; I²=89%, P<0.001) than the control group.

Figure 2. Effect of exercise stress testing and abrupt stand testing on QTc interval.

A, QTc interval at baseline; (B) ΔQTc at peak exercise; (C) ΔQTc at recovery 4 to 5 minutes; (D) ΔQTc at abrupt stand. LQTS indicates long QT syndrome; and SD, standardized difference.

Effect of Exercise Stress Testing on QTc Interval in Different Subtypes of LQTS

To facilitate genotyping of LQTS, we performed a subgroup analysis of the 3 most common LQTS (LQT1, LQT2, and LQT3). As shown in Figure 3A and 3B, ΔQTc was nearly 40 ms longer both at peak exercise (MD, 36.72 [95% CI, 33.08–40.36]; P<0.001) and recovery 4 to 5 minutes (MD, 39.72 [95% CI, 35.06–44.38]; P<0.001) in patients with LQT1 than control. In contrast, ΔQTc were shorter at peak exercise (MD, −8.84 [95% CI, −12.62 to −5.07]; P<0.001), but longer at recovery 4 to 5 minutes (MD, 11.69 [95% CI, 6.48–16.91]; P<0.001) in patients with LQT2 (Figure 3C and 3D). Unlike subjects with LQT1 or LQT2, ΔQTc in patients with LQT3 was both shorter at peak exercise (MD, −14.06 [95% CI, −29.02 to 0.89]; I²=8%, P=0.07), and recovery at 4 to 5 minutes (MD, −15.84 [95% CI, −28.10 to −3.58]; I²=90%, P<0.05), which was confined to a limited population. Results are shown in Figure 3E and 3F.

Figure 3. Effect of exercise stress testing on QTc interval in different subtypes of LQTS.

A, ΔQTc of LQT1 at peak exercise; (B) ΔQTc of LQT1 at recovery 4 to 5 minutes; (C) ΔQTc of LQT2 at peak exercise; (D) ΔQTc of LQT2 at recovery 4 to 5 minutes; (E) ΔQTc of LQT3 at peak exercise; (F) ΔQTc of LQT3 at recovery 4 to 5 minutes. LQTS indicates long QT syndrome; and SD, standardized difference.

Effect of Epinephrine Infusion on QTc Interval

With epinephrine infusion, the QTc interval was prolonged both in control and patients with LQTS, while patients with LQTS had 70 ms longer ΔQTc than control (MD, 70.16 [95% CI, 58.80–81.52]; I²=3%, P<0.001), as shown in Figure 4A. In terms of the most common LQTS subtypes, patients with both LQT1 and LQT2 had longer ΔQTc than control, more obviously in LQT1 (MD, 68.26 [95% CI, 58.91–77.60]; P<0.001) than LQT2 (MD, 60.17 [95% CI, 50.18–70.16]; P<0.001). On the contrary, in patients with LTQ3 with epinephrine infusion, ΔQTc were remarkably shorter than control (MD, −17.36 [95% CI, −30.01 to −4.63]; I²=35%, P<0.001), which was consistent with exercise stress testing, as shown in Figure 4B through 4D.

Figure 4. Effect of epinephrine infusion on QTc interval.

A, ΔQTc of LQTS in epinephrine infusion test; (B) ΔQTc of LQT1 in epinephrine infusion test; (C) ΔQTc of LQT2 in epinephrine infusion test; (D) ΔQTc of LQT3 in epinephrine infusion test. LQTS indicates long QT syndrome; and SD, standardized difference.

Trend of ΔQTc in Provocative Testing

As Figure 5A shows, in the control population, abrupt stand prolonged QTc interval by 40 ms; in the subsequent exercise testing, either at peak exercise or recovery 4 to 5 minutes, QTc interval was consistent with baseline. As Figure 5B shows, in the population with LQTS, abrupt stand prolonged QTc interval by nearly 70 ms, while QTc interval at peak exercise or recovery 4 to 5 minutes were prolonged mildly.

Figure 5. The trend of ΔQTc in provocative testing.

A through E, ΔQTc in abrupt stand testing and exercise stress testing. A, Control, (B) LQTS, (C) LQT1, (D) LQT2, (E) LQT3; (F) ΔQTc in epinephrine QT stress testing. Red asterisk represents the composite means of studies, and dashed line represents the trend of ΔQTc. LQTS indicates long QT syndrome.

To clarify the manifestations of LQTS subtypes, we analyzed LQT1, LQT2, and LQT3, separately. In patients with LQT1, abrupt stand prolonged QTc interval by 40 ms, the QTc interval was also prolonged by 40 ms at peak exercise, and 30 ms at recovery 4 to 5 minutes (Figure 5C). In patients with LQT2, abrupt stand prolonged QTc interval by nearly 50 ms, and QTc interval was nearly consistent with baseline at peak exercise and recovery 4 to 5 minutes (Figure 5D). In patients with LQT3, QTc interval was both shortened at peak exercise and recovery 4 to 5 minutes (Figure 5E). Unfortunately, no data were available on QTc interval by abrupt stand in patients with LQT3. Moreover, QTc interval was prolonged in both control and patients with LQTS with epinephrine infusion, especially in patients with LQT1 and LQT2 (Figure 5F).

Diagnostic Workflow for Diagnosis and Genotyping of LQTS

Based on the current meta‐analysis, we combined baseline electrocardiogram (ECG) (QTc interval) as the first step, abrupt stand testing plus exercise testing (QTc trend) as the second step, and epinephrine testing (ΔQTc) as the third step to diagnosis and genotyping of LQTS (Figure 6A). First, based on QTc interval in baseline ECG, overt LQTS was distinguished from concealed LQTS or normal population, because QTc ≥480 ms in women or ≥470 ms in men was determined with 100% specificity for the diagnosis of LQTS. ⁴⁵ While in family members with LQTS, considering a priori likelihood, QTc ≥430 ms can be used as a cutoff to distinguish carriers from noncarriers. ⁴⁶ Then, QTc trend in abrupt stand testing plus exercise testing (treadmill or bicycle) were used to genotype LQTS, in which the QTc trend exhibited 3 types, named Type Ⅰ, Type Ⅱ, and Type Ⅲ. In Type Ⅰ (Figure 6B), the QTc interval increased markedly in abrupt stand and peak exercise, and persists despite HR recovery during the recovery period, seen in patients with LQT1; in type Ⅱ (Figure 6C), the QTc interval increased markedly in the abrupt stand, shortened slightly in peak exercise, and lengthened mildly during the recovery period, seen in patients with LQT2 or normal population, while in type Ⅲ (Figure 6D), the QTc interval shortened during the entire test, and persists during the recovery period despite HR recovery, seen in patients with LQT3. Diagnosis of LQTS was particularly challenging in patients with normal or borderline QTc interval in baseline ECG and presented type Ⅱ QTc trend in abrupt stand testing plus exercise testing. Consequently, we propose that an epinephrine testing be performed in patients suspected to have LQTS, in whom larger ΔQTc would suggest LQT2, and smaller ΔQTc suggests unaffected individuals.

Figure 6. Diagnostic workflow of LQTS.

A, Diagnostic workflow of LQTS; (B) Type Ⅰ QTc trend; (C) Type Ⅱ QTc trend; (D) Type Ⅲ QTc trend. ECG indicates electrocardiogram; and LQTS, long QT syndrome.

DISCUSSION

To the best of our knowledge, this is the first attempt to determine the pooled utility of provocative testing in the diagnosis and genotyping of LQTS. The main findings of the present meta‐analysis are as follows: (1) Insufficient QT interval shortening in response to tachycardia, delayed QT interval recovery at the late recovery of exercise, is of great value for distinguishing LQTS from control. (2) QTc interval response to abrupt stand testing and exercise testing differed between LQTS subtypes (Type Ⅰ, Type Ⅱ, and Type Ⅲ), which is of additional value for genotyping LQTS. (3) Epinephrine infusion testing prolonged QTc interval both in control and patients with LQTS, most obviously in patients with LQT1 and LQT2, which is a potential powerful test for diagnosis and genotyping of LQTS. (4) We first draft a diagnostic workflow of LQTS, which may be helpful for the diagnosis and genotyping of LQTS, but still deserves further evaluation.

Based on the classic Schwartz score, diagnosis of LQTS is relatively straightforward in patients with overt QT prolongation or clinical symptoms. However, 25% to 50% of LQTS carriers have a normal or borderline resting QTc (concealed LQTS) because of a combination of variable penetrance, the effect of genes modification, and the dynamic nature of QT interval. ⁵ Furthermore, up to 40% of individuals with congenital LQTS are “silent mutation carriers” (asymptomatic LQTS), who are exposed to a 10 times higher risk of CEs than unaffected family members. ⁵ , ⁴⁷ Thus, accurate identification of concealed LQTS and asymptomatic LQTS is of great importance, despite being particularly challenging. Genetic testing has been regarded as the criterion standard for diagnosis and genotyping of LQTS, but it is restricted by its cost and time delay. Consequently, several studies tried to evaluate additional readily applicable clinical tests, including abrupt stand testing, ¹⁹ treadmill test, ¹³ bicycle test, ¹⁸ epinephrine QT stress testing, ²⁶ adenosine exposure, ³⁸ mental stress, ¹⁸ Valsalva maneuver, ⁴⁰ and cold‐water face immersion. ⁷

Sympathetic activation has been considered an important trigger of malignant arrhythmia in patients with LQTS, and β‐blockers or left cardiac sympathetic denervation have been recommended to reduce the occurrence of CEs, ¹ indicating that the autonomic nervous (sympathetic and parasympathetic nervous) system may influence the electrophysiological properties of the myocardium. Studies have shown that provocative testing, including the most common abrupt stand testing, exercise testing (treadmill and bicycle), and epinephrine infusion testing, plays an important role in diagnosis and risk stratification of LQTS. ⁵ , ¹⁶ , ¹⁹ , ²⁴ , ⁴⁸ In terms of provocative tests, treadmill or bicycle tests are accessible in most pediatric and adult cardiology centers, both with a variety of protocols (gradual or burst increase), and can be recorded with ECG in supine position, upon standing, and any time during exercise. ¹⁰ Studies showed that QTc upon standing, peak exercise, and late recovery are more sensitive markers of genotype‐specific repolarization responses than other times during exercise. ⁵ , ⁸

In our meta‐analysis, we found that LQT1‐3 corresponds to 3 types of QTc trend during abrupt stand testing plus exercise testing, LQT1 presented increased QTc upon standing and peak exercise, and persists in late recovery (type Ⅰ); LQT3 presented shortened QTc during the entire test (Type Ⅲ), while LQT2 was similar to control, QTc increased upon standing, and returned to baseline or mildly shortened at peak exercise, and gradually increased in late recovery (Type Ⅱ). The patterns of Type Ⅰ and Type Ⅱ had been found in previous studies, ¹⁰ , ¹⁴ , ¹⁵ , ¹⁶ but without graphic description. Despite patients with LQT2 having 9 ms shorter ΔQTc at peak exercise, and 12 ms longer ΔQTc at late recovery than control, these differences were not enough to distinguish patients with LQT2 from control, especially concealed LQT2. Furthermore, given the relative rarity of LQT3, our analysis enrolled only 74 patients with LQT3; thus the QTc trend of Type Ⅲ deserves further evaluation.

In contrast, the epinephrine infusion testing, whether by bolus infusion (Shimizu protocol) or an incremental infusion (Mayo protocol), should be performed at LQTS specialty centers and requires a period of β‐blockers washout, which were not needed in exercise testing. ¹⁰ Magnano ⁴⁹ found that QTc was prolonged by epinephrine (79±40 ms) in healthy subjects. In our meta‐analysis, epinephrine infusion testing prolonged QTc both in control and patients with LQTS, most obviously in LQT1 and LQT2, up to 60 ms difference of ΔQTc between patients with LQT1/LQT2 and control, which is a meaningful marker for distinguishing LQT1/LQT2 from control, covering the shortage of exercise testing. Although all of these studies provided a meaningful theoretical basis, the appropriate cut‐off of ΔQTc is still unknown. ²³ , ²⁴ , ²⁵ Furthermore, in the Shimizu protocol, electrocardiography data were collected under baseline and at steady‐state conditions of epinephrine ²⁶ ; while in the Mayo protocol, ΔQTc were calculated by the difference between the maximal and minimal QTc. ²⁴ Whether the range of ΔQTc is consistent between different protocols is still unknown. Because of its arrhythmia risk and controversial evidence, guidelines recommend limited levels of epinephrine testing in the evaluation of a cardiac arrest survivor (class 2b). ⁵⁰ Therefore, in our diagnostic workflow, epinephrine testing is proposed for use in patients with normal or borderline QTc in baseline ECG and presented Type Ⅱ QTc trend in exercise testing, in which patients with overt ΔQTc were probably concealed LQT2, otherwise, were normal population, which deserve further genetic testing validation.

Limitations

Given the nature of our study as a systematic review and meta‐analysis, our analyses are limited by the reported data in the original reports. Limitations of our meta‐analysis are as follows: (1) Given the relative rarity of LQTS, we only included 1137 patients with LQTS in our analysis; these mainly included patients with LQT1 and LQT2, and analysis enrolled only 74 patients with LQT3, so the pattern of Type Ⅲ QTc trend warrants further evaluation. (2) The data were presented differently between studies, restricting analysis of optimal cut‐off, and corresponding sensitivity and specificity, especially epinephrine infusion testing, making our diagnostic workflow of limited diagnostic value. (3) Fewer studies analyzed concealed or asymptomatic LQTS separately, restricting analysis of the diagnostic value of the provocative tests in concealed or asymptomatic LQTS. (4) Patients with LQTS have a changed risk of CEs as time goes on, so the different reactions of provocative testing between adolescents and children remain unknown. (5) Most of our included studies were retrospective cohort, only with derivation cohort, and lacked external validation, which inherently limited the quality of evidence. (6) Our meta‐analysis has high heterogeneity between studies, mainly because studies enrolled different subtypes of LQTS, different types of provocative tests, and differences in QTc measurement. Given the above, there is an urgent need for prospective, multicenter, large‐scale trials to validate the diagnostic value of our diagnostic workflow, especially in concealed or asymptomatic LQTS.

CONCLUSIONS

QTc trend Type Ⅰ and Type Ⅲ during abrupt stand testing and exercise testing can be used to propose a prospective evaluation of LQT1 and LQT3, respectively. Type Ⅱ QTc trend combined epinephrine infusion testing could distinguish LQT2 from control. A preliminary diagnostic workflow was drawn up, but deserves further evaluation.

Sources of Funding

This work was supported by Beijing Municipal Science & Technology Commission (No. Z191100006619007) and National Natural Science Foundation of China (No. 82100333).

Disclosures

None.