Determination and Interpretation of the QT Interval: Comprehensive Analysis of a Large Cohort of Long QT Syndrome Patients and Controls

Arja Suzanne Vink, Benjamin Neumann, Krystien V.V. Lieve, Moritz F. Sinner, Nynke Hofman, Soufiane El Kadi, Melissa H.A. Schoenmaker, Hanneke M.J. Slaghekke, Jonas S.S.G. De Jong, Sally Ann B. Clur, Nico A. Blom, Stefan Kääb, Arthur A.M. Wilde, Pieter G. Postema

Research output: Contribution to journalArticleAcademicpeer-review

89 Citations (Scopus)


Background: Long QT syndrome (LQTS) is associated with potentially fatal arrhythmias. Treatment is very effective, but its diagnosis may be challenging. Importantly, different methods are used to assess the QT interval, which makes its recognition difficult. QT experts advocate manual measurements with the tangent or threshold method. However, differences between these methods and their performance in LQTS diagnosis have not been established. We aimed to assess similarities and differences between these 2 methods for QT interval analysis to aid in accurate QT assessment for LQTS. Methods: Patients with a confirmed pathogenic variant in KCNQ1(LQT1), KCNH2(LQT2), or SCN5A(LQT3) genes and their family members were included. Genotype-positive patients were identified as LQTS cases and genotype-negative family members as controls. ECGs were analyzed with both methods, providing inter- and intrareader validity and diagnostic accuracy. Cutoff values based on control population's 95th and 99th percentiles, and LQTS-patients' 1st and 5th percentiles were established based on the method to correct for heart rate, age, and sex. Results: We included 1484 individuals from 265 families, aged 33±21 years and 55% females. In the total cohort, QTTangent was 10.4 ms shorter compared with QTThreshold (95% limits of agreement±20.5 ms, P<0.0001). For all genotypes, QTTangent was shorter than QTThreshold (P<0.0001), but this was less pronounced in LQT2. Both methods yielded a high inter- and intrareader validity (intraclass correlation coefficient >0.96), and a high diagnostic accuracy (area under the curve >0.84). Using the current guideline cutoff (QTc interval 480 ms), both methods had similar specificity but yielded a different sensitivity. QTc interval cutoff values of QTTangent were lower compared with QTThreshold and different depending on the correction for heart rate, age, and sex. Conclusion: The QT interval varies depending on the method used for its assessment, yet both methods have a high validity and can both be used in diagnosing LQTS. However, for diagnostic purposes current guideline cutoff values yield different results for these 2 methods and could result in inappropriate reassurance or treatment. Adjusted cutoff values are therefore specified for method, correction formula, age, and sex. In addition, a freely accessible online probability calculator for LQTS ( has been made available as an aid in the interpretation of the QT interval.

Original languageEnglish
Pages (from-to)2345-2358
Number of pages14
Issue number21
Publication statusPublished - 20 Nov 2018


  • Adolescent
  • Adult
  • Age Factors
  • Case-Control Studies
  • Child
  • Databases, Factual
  • ERG1 Potassium Channel/genetics
  • Electrocardiography
  • Female
  • Genotype
  • Humans
  • KCNQ1 Potassium Channel/genetics
  • Long QT Syndrome/diagnosis
  • Male
  • Middle Aged
  • NAV1.5 Voltage-Gated Sodium Channel/genetics
  • Retrospective Studies
  • Sex Factors
  • Young Adult

Cite this