Universal Correction for QT/RR Hysteresis

Heart rate stability in a clinical setting and after a short exercise in healthy male volunteers

INTRODUCTION

Limited data exist on heart rate stabilization in the domiciled nature of phase I clinical studies, particularly when frequent measurements of QT intervals are involved. The present analysis aimed to evaluate heart rate stability in the domiciled nature of, and stabilization after a short exercise.

METHODS

Fifty-six healthy male subjects were included in this analysis. Data during a domiciled clinical setting and after a short exercise were analysed. Mean values of 30 s intervals of collected electrocardiographical data (PR, RR, QT and QTcF intervals) during a 10-min supine resting period in a domiciled nature or after walking up and down three stories (100 steps) were compared to baseline values using paired t-tests or compared to the intrasubject standard deviation.

RESULTS

Stable heart rates and stable QTcF intervals observed immediately upon assuming a supine position in the domiciled clinical setting. After the short exercise, PR interval and RR interval were significantly (p < 0.05) shorter for up to 120 s (mean value -9.8 ± 7.2 ms) and 30 s (-160 ± 165 ms, p < 0.05), respectively. QT and QTcF intervals were significantly (p < 0.05) shorter for up to 90 and 120 s postexercise, respectively. Both QT and QTcF intervals stabilized after 2 min, but QT interval remained prolonged while QTcF interval returned to baseline levels.

CONCLUSION

In a clinical setting, male volunteers do not require a waiting period for electrocardiographic parameter normalization. However, accurate measurement of these parameters following a short exercise necessitates a minimum 2-min resting interval.

Accurate QT correction method from transfer entropy

Background

The QT interval in the electrocardiogram (ECG) is a fundamental risk measure for arrhythmic adverse cardiac events. However, the QT interval depends on the heart rate and must be corrected accordingly. The present QT correction (QTc) methods are either simple models leading to under- or overcorrection, or impractical in requiring long-term empirical data. In general, there is no consensus on the best QTc method.

Objective

We introduce a model-free QTc method-AccuQT-that computes QTc by minimizing the information transfer from R-R to QT intervals. The objective is to establish and validate a QTc method that provides superior stability and reliability without models or empirical data.

Methods

We tested AccuQT against the most commonly used QT correction methods by using long-term ECG recordings of more than 200 healthy subjects from PhysioNet and THEW databases.

Results

AccuQT overperforms the previously reported correction methods: the proportion of false-positives is reduced from 16% (Bazett) to 3% (AccuQT) for the PhysioNet data. In particular, the QTc variance is significantly reduced and thus the RR-QT stability is increased.

Conclusion

AccuQT has significant potential to become the QTc method of choice in clinical studies and drug development. The method can be implemented in any device recording R-R and QT intervals.

Intra-subject stability of different expressions of spatial QRS-T angle and their relationship to heart rate

Three-dimensional angle between the QRS complex and T wave vectors is a known powerful cardiovascular risk predictor. Nevertheless, several physiological properties of the angle are unknown or poorly understood. These include, among others, intra-subject profiles and stability of the angle relationship to heart rate, characteristics of angle/heart-rate hysteresis, and the changes of these characteristics with different modes of QRS-T angle calculation. These characteristics were investigated in long-term 12-lead Holter recordings of 523 healthy volunteers (259 females). Three different algorithmic methods for the angle computation were based on maximal vector magnitude of QRS and T wave loops, areas under the QRS complex and T wave curvatures in orthogonal leads, and weighted integration of all QRS and T wave vectors moving around the respective 3-dimensional loops. These methods were applied to orthogonal leads derived either by a uniform conversion matrix or by singular value decomposition (SVD) of the original 12-lead ECG, giving 6 possible ways of expressing the angle. Heart rate hysteresis was assessed using the exponential decay models. All these methods were used to measure the angle in 659,313 representative waveforms of individual 10-s ECG samples and in 7,350,733 individual beats contained in the same 10-s samples. With all measurement methods, the measured angles fitted second-degree polynomial regressions to the underlying heart rate. Independent of the measurement method, the angles were found significantly narrower in females (p &lt; 0.00001) with the differences to males between 10o and 20o, suggesting that in future risk-assessment studies, different angle dichotomies are needed for both sexes. The integrative method combined with SVD leads showed the highest intra-subject reproducibility (p &lt; 0.00001). No reproducible delay between heart rate changes and QRS-T angle changes was found. This was interpreted as a suggestion that the measurement of QRS-T angle might offer direct assessment of cardiac autonomic responsiveness at the ventricular level.

Short-Term Beat-to-Beat QT Variability Appears Influenced More Strongly by Recording Quality Than by Beat-to-Beat RR Variability

Increases in beat-to-beat variability of electrocardiographic QT interval duration have repeatedly been associated with increased risk of cardiovascular events and complications. The measurements of QT variability are frequently normalized for the underlying RR interval variability. Such normalization supports the concept of the so-called immediate RR effect which relates each QT interval to the preceding RR interval. The validity of this concept was investigated in the present study together with the analysis of the influence of electrocardiographic morphological stability on QT variability measurements. The analyses involved QT and RR measurements in 6,114,562 individual beats of 642,708 separate 10-s ECG samples recorded in 523 healthy volunteers (259 females). Only beats with high morphology correlation (r &gt; 0.99) with representative waveforms of the 10-s ECG samples were analyzed, assuring that only good quality recordings were included. In addition to these high correlations, SDs of the ECG signal difference between representative waveforms and individual beats expressed morphological instability and ECG noise. In the intra-subject analyses of both individual beats and of 10-s averages, QT interval variability was substantially more strongly related to the ECG noise than to the underlying RR variability. In approximately one-third of the analyzed ECG beats, the prolongation or shortening of the preceding RR interval was followed by the opposite change of the QT interval. In linear regression analyses, underlying RR variability within each 10-s ECG sample explained only 5.7 and 11.1% of QT interval variability in females and males, respectively. On the contrary, the underlying ECG noise contents of the 10-s samples explained 56.5 and 60.1% of the QT interval variability in females and males, respectively. The study concludes that the concept of stable and uniform immediate RR interval effect on the duration of subsequent QT interval duration is highly questionable. Even if only stable beat-to-beat measurements of QT interval are used, the QT interval variability is still substantially influenced by morphological variability and noise pollution of the source ECG recordings. Even when good quality recordings are used, noise contents of the electrocardiograms should be objectively examined in future studies of QT interval variability.

Sex and Rate Change Differences in QT/RR Hysteresis in Healthy Subjects

While it is now well-understood that the extent of QT interval changes due to underlying heart rate differences (i.e., the QT/RR adaptation) needs to be distinguished from the speed with which the QT interval reacts to heart rate changes (i.e., the so-called QT/RR hysteresis), gaps still exist in the physiologic understanding of QT/RR hysteresis processes. This study was designed to address the questions of whether the speed of QT adaptation to heart rate changes is driven by time or by number of cardiac cycles; whether QT interval adaptation speed is the same when heart rate accelerates and decelerates; and whether the characteristics of QT/RR hysteresis are related to age and sex. The study evaluated 897,570 measurements of QT intervals together with their 5-min histories of preceding RR intervals, all recorded in 751 healthy volunteers (336 females) aged 34.3 ± 9.5 years. Three different QT/RR adaptation models were combined with exponential decay models that distinguished time-based and interval-based QT/RR hysteresis. In each subject and for each modelling combination, a best-fit combination of modelling parameters was obtained by seeking minimal regression residuals. The results showed that the response of QT/RR hysteresis appears to be driven by absolute time rather than by the number of cardiac cycles. The speed of QT/RR hysteresis was found decreasing with increasing age whilst the duration of individually rate corrected QTc interval was found increasing with increasing age. Contrary to the longer QTc intervals, QT/RR hysteresis speed was faster in females. QT/RR hysteresis differences between heart rate acceleration and deceleration were not found to be physiologically systematic (i.e., they differed among different healthy subjects), but on average, QT/RR hysteresis speed was found slower after heart rate acceleration than after rate deceleration.

The best QT correction formula in a non-hospitalized population: the Fasa PERSIAN cohort study

Background

QT interval as an indicator of ventricular repolarization is a clinically important parameter on an electrocardiogram (ECG). QT prolongation predisposes individuals to different ventricular arrhythmias and sudden cardiac death. The current study aimed to identify the best heart rate corrected QT interval for a non-hospitalized Iranian population based on cardiovascular mortality.

Methods

Using Fasa PERSIAN cohort study data, this study enrolled 7071 subjects aged 35–70 years. Corrected QT intervals (QTc) were calculated by the QT interval measured by Cardiax® software from ECGs and 6 different correction formulas (Bazett, Fridericia, Dmitrienko, Framingham, Hodges, and Rautaharju). Mortality status was checked using an annual telephone-based follow-up and a minimum 3-year follow-up for each participant. Bland–Altman, QTc/RR regression, sensitivity analysis, and Cox regression were performed in IBM SPSS Statistics v23 to find the best QT. Also, for calculating the upper and lower limits of normal of different QT correction formulas, 3952 healthy subjects were selected.

Results

In this study, 56.4% of participants were female, and the mean age was 48.60 ± 9.35 years. Age, heart rate in females, and QT interval in males were significantly higher. The smallest slopes of QTc/RR analysis were related to Fridericia in males and Rautaharju followed by Fridericia in females. Thus, Fridericia’s formula was identified as the best mathematical formula and Bazett’s as the worst in males. In the sensitivity analysis, however, Bazett’s formula had the highest sensitivity (23.07%) among all others in cardiac mortality. Also, in the Cox regression analysis, Bazett’s formula was better than Fridericia’s and was identified as the best significant cardiac mortality predictor (Hazard ratio: 4.31, 95% CI 1.73–10.74, p value = 0.002).

Conclusion

Fridericia was the best correction formula based on mathematical methods. Bazett’s formula despite its poorest performance in mathematical methods, was the best one for cardiac mortality prediction. Practically, it is suggested that physicians use QTcB for a better evaluation of cardiac mortality risk. However, in population-based studies, QTcFri might be the one to be used by researchers.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12872-022-02502-2.

Influence of heart rate correction formulas on QTc interval stability

Monitoring of QTc interval is mandated in different clinical conditions. Nevertheless, intra-subject variability of QTc intervals reduces the clinical utility of QTc monitoring strategies. Since this variability is partly related to QT heart rate correction, 10 different heart rate corrections (Bazett, Fridericia, Dmitrienko, Framingham, Schlamowitz, Hodges, Ashman, Rautaharju, Sarma, and Rabkin) were applied to 452,440 ECG measurements made in 539 healthy volunteers (259 females, mean age 33.3 ± 8.4 years). For each correction formula, the short term (5-min time-points) and long-term (day-time hours) variability of rate corrected QT values (QTc) was investigated together with the comparisons of the QTc values with individually corrected QTcI values obtained by subject-specific modelling of the QT/RR relationship and hysteresis. The results showed that (a) both in terms of short-term and long-term QTc variability, Bazett correction led to QTc values that were more variable than the results of other corrections (p < 0.00001 for all), (b) the QTc variability by Fridericia and Framingham corrections were not systematically different from each other but were lower than the results of other corrections (p-value between 0.033 and < 0.00001), and (c) on average, Bazett QTc values departed from QTcI intervals more than the QTc values of other corrections. The study concludes that (a) previous suggestions that Bazett correction should no longer be used in clinical practice are fully justified, (b) replacing Bazett correction with Fridericia and/or Framingham corrections would improve clinical QTc monitoring, (c) heart rate stability is needed for valid QTc assessment, and (d) development of further QTc corrections for day-to-day use is not warranted.

Spatial distribution of physiologic 12-lead QRS complex

The normal physiologic range of QRS complex duration spans between 80 and 125 ms with known differences between females and males which cannot be explained by the anatomical variations of heart sizes. To investigate the reasons for the sex differences as well as for the wide range of normal values, a technology is proposed based on the singular value decomposition and on the separation of different orthogonal components of the QRS complex. This allows classification of the proportions of different components representing the 3-dimensional representation of the electrocardiographic signal as well as classification of components that go beyond the 3-dimensional representation and that correspond to the degree of intricate convolutions of the depolarisation sequence. The technology was applied to 382,019 individual 10-s ECG samples recorded in 639 healthy subjects (311 females and 328 males) aged 33.8 ± 9.4 years. The analyses showed that QRS duration was mainly influenced by the proportions of the first two orthogonal components of the QRS complex. The first component demonstrated statistically significantly larger proportion of the total QRS power (expressed by the absolute area of the complex in all independent ECG leads) in females than in males (64.2 ± 11.6% vs 59.7 ± 11.9%, p < 0.00001—measured at resting heart rate of 60 beats per minute) while the second component demonstrated larger proportion of the QRS power in males compared to females (33.1 ± 11.9% vs 29.6 ± 11.4%, p < 0.001). The analysis also showed that the components attributable to localised depolarisation sequence abnormalities were significantly larger in males compared to females (2.85 ± 1.08% vs 2.42 ± 0.87%, p < 0.00001). In addition to the demonstration of the technology, the study concludes that the detailed convolution of the depolarisation waveform is individual, and that smoother and less intricate depolarisation propagation is the mechanism likely responsible for shorter QRS duration in females.

PK/PD modeling of a clazosentan thorough QT study with hysteresis in concentration-QT and RR-QT

Clazosentan's potential QT liability was investigated in a thorough QT study in which clazosentan was administered intravenously as a continuous infusion of 20 mg/h immediately followed by 60 mg/h. Clazosentan prolonged the placebo-corrected change-from-baseline QT interval corrected for RR with Fridericia's formula (ΔΔQTcF) with the maximum QT effect occurring 4 h after the maximum drug concentration, apparently associated with vomiting. The delayed effect precluded the standard linear modeling approach. This analysis aimed at characterizing the concentration-QT relationship in consideration of RR-QT hysteresis, concentration-ΔΔQTcF hysteresis, and the influence of vomiting. Nonlinear mixed-effects modeling was applied to characterize pharmacokinetics and pharmacodynamics, i.e., ΔΔQTcF. Simulations were used to predict ΔΔQTcF for expected therapeutic dose used in Phase 3 clinical development. Correction for RR-QT hysteresis did not influence ΔΔQTcF to a relevant extent. Pharmacokinetics of clazosentan were best described by a linear two-compartment model. The delayed QT prolongation was characterized by an indirect-response model with loglinear drug effect. Vomiting had no statistically significant influence on QT prolongation despite apparent differences between subjects vomiting and not vomiting, probably since vomiting occurred mostly after the main QT prolongation. Following a simulated 3-h infusion of 15 mg/h of clazosentan, the upper bound of the predicted 90% CI for mean ΔΔQTcF was expected to exceed the 10-ms regulatory threshold of concern with maximum effect 3.5 h after end of infusion. TRN: NCT03657446, 05 Sep 2018.

Heart Rate Dependency and Inter-Lead Variability of the T Peak - T End Intervals

The electrocardiographic (ECG) assessment of the T peak–T end (Tpe) intervals has been used in many clinical studies, but several related physiological aspects have not been reported. Specifically, the sources of the Tpe differences between different ECG leads have not been systematically researched, the relationship of Tpe duration to underlying heart rate has not been firmly established, and little is known about the mutual correspondence of Tpe intervals measured in different ECG leads. This study evaluated 796,620 10-s 12-lead ECGs obtained from long-term Holters recorded in 639 healthy subjects (311 female) aged 33.8 ± 9.4 years. For each ECG, transformation to orthogonal XYZ lead was used to measure Tpe in the orthogonal vector magnitude (used as a reference for lead-to-lead comparisons) and to construct a three-dimensional T wave loop. The loop roundness was expressed by a ratio between its circumference and length. These ratios were significantly related to the standard deviation of Tpe durations in different ECG leads. At the underlying heart rate of 60 beats per minute, Tpe intervals were shorter in female than in male individuals (82.5 ± 5.6 vs 90.0 ± 6.5 ms, p < 0.0001). When studying linear slopes between Tpe intervals measured in different leads and the underlying heart rate, we found only minimal heart rate dependency, which was not systematic across the ECG leads and/or across the population. For any ECG lead, positive Tpe/RR slope was found in some subjects (e.g., 79 and 25% of subjects for V2 and V4 measurements, respectively) and a negative Tpe/RR slope in other subjects (e.g., 40 and 65% for V6 and V5, respectively). The steepest positive and negative Tpe/RR slopes were found for measurements in lead V2 and V4, respectively. In all leads, the Tpe/RR slope values were close to zero, indicating, on average, Tpe changes well below 2 ms for RR interval changes of 100 ms. On average, longest Tpe intervals were measured in lead V2, the shortest in lead III. The study concludes that the Tpe intervals measured in different leads cannot be combined. Irrespective of the measured ECG lead, the Tpe interval is not systematically heart rate dependent, and no heart rate correction should be used in clinical Tpe investigations.

Problems with Bazett QTc correction in paediatric screening of prolonged QTc interval

Background

Bazett formula is frequently used in paediatric screening for the long QT syndrome (LQTS) and proposals exist that using standing rather than supine electrocardiograms (ECG) improves the sensitivity of LQTS diagnosis. Nevertheless, compared to adults, children have higher heart rates (especially during postural provocations) and Bazett correction is also known to lead to artificially prolonged QTc values at increased heart rates. This study assessed the incidence of erroneously increased QTc values in normal children without QT abnormalities.

Methods

Continuous 12-lead ECGs were recorded in 332 healthy children (166 girls) aged 10.7 ± 2.6 years while they performed postural manoeuvring consisting of episodes (in the following order) of supine, sitting, standing, supine, standing, sitting, and supine positions, each lasting 10 min. Detailed analyses of QT/RR profiles confirmed the absence of prolonged individually corrected QTc interval in each child. Heart rate and QT intervals were measured in 10-s ECG segments and in each segment, QTc intervals were obtained using Bazett, Fridericia, and Framingham formulas. In each child, the heart rates and QTc values obtained during supine, sitting and standing positions were averaged. QTc durations by the three formulas were classified to < 440 ms, 440–460 ms, 460–480 ms, and > 480 ms.

Results

At supine position, averaged heart rate was 77.5 ± 10.5 beat per minute (bpm) and Bazett, Fridericia and Framingham QTc intervals were 425.3 ± 15.8, 407.8 ± 13.9, and 408.2 ± 13.1 ms, respectively. At sitting and standing, averaged heart rate increased to 90.9 ± 10.1 and 100.9 ± 10.5 bpm, respectively. While Fridericia and Framingham formulas showed only minimal QTc changes, Bazett correction led to QTc increases to 435 ± 15.1 and 444.9 ± 15.9 ms at sitting and standing, respectively. At sitting, Bazett correction identified 51, 4, and 0 children as having the QTc intervals 440–460, 460–480, and > 480 ms, respectively. At sitting, these numbers increased to 118, 11, and 1, while on standing these numbers were 151, 45, and 5, respectively. Irrespective of the postural position, Fridericia and Framingham formulas identified only a small number (< 7) of children with QT interval between 440 and 460 ms and no children with longer QTc.

Conclusion

During screening for LQTS in children, the use of Bazett formula leads to a high number of false positive cases especially if the heart rates are increased (e.g. by postural manoeuvring). The use of Fridericia formula can be recommended to replace the Bazett correction not only for adult but also for paediatric ECGs.

Potential strategy for assessing QT/QTc interval for drugs that produce rapid changes in heart rate: Electrocardiographic assessment of the effects of intravenous remimazolam on cardiac repolarization

Aims

Remimazolam is a new, ultra‐short‐acting benzodiazepine developed for intravenous (IV) use during procedural sedation and in general anaesthesia. Two trials were conducted to characterize its effects on cardiac repolarization.

Methods

A thorough QT/QTc (TQT) study assessed electrocardiography effects of therapeutic and supratherapeutic doses of remimazolam and midazolam. To investigate whether RR‐QT hysteresis effects due to rapid heart rate changes might have confounded the QTc assessments in the TQT trial, a second trial used continuous IV remimazolam infusion to achieve stable heart rates during periods of stable remimazolam plasma concentration.

Results

During the TQT, both compounds produced a 10–20‐beats/min increase in heart rate within 30 seconds, persisting for 5–10 minutes. Within 30 seconds, the upper bound of the 2‐sided 90% confidence interval for the placebo‐corrected change from baseline for QTcI (ΔΔQTcI) exceeded 10 ms for both doses of remimazolam (ΔΔQTcI 7.2 [3.2, 11.2] ms for the 10 mg dose and 10.4 [6.5, 14.3] ms for the 20 mg dose) as well as for the 7.5‐mg dose of midazolam (8.2 [4.4, 12.1] ms). At 2 minutes after IV bolus, the upper bound of the 2‐sided 90% confidence interval for ΔΔQTcI exceeded 10 ms only for the remimazolam 20‐mg dose (6.3 [2.3, 10.2] ms). During the second study, during periods of stable heart rate, remimazolam had no clinically significant effect on QTc (peak ΔΔQTcI 3.4 [−1.1, 7.6] ms).

Conclusion

Remimazolam does not prolong cardiac repolarization (QTc). The methods reported here may allow assessment of the QTc effects of other drugs given by IV bolus.

Physiologic heart rate dependency of the PQ interval and its sex differences

On standard electrocardiogram (ECG) PQ interval is known to be moderately heart rate dependent, but no physiologic details of this dependency have been established. At the same time, PQ dynamics is a clear candidate for non-invasive assessment of atrial abnormalities including the risk of atrial fibrillation. We studied PQ heart rate dependency in 599 healthy subjects (aged 33.5 ± 9.3 years, 288 females) in whom drug-free day-time 12-lead ECG Holters were available. Of these, 752,517 ECG samples were selected (1256 ± 244 per subject) to measure PQ and QT intervals and P wave durations. For each measured ECG sample, 5-minute history of preceding cardiac cycles was also obtained. Although less rate dependent than the QT intervals (36 ± 19% of linear slopes), PQ intervals were found to be dependent on underlying cycle length in a highly curvilinear fashion with the dependency significantly more curved in females compared to males. The PQ interval also responded to the heart rate changes with a delay which was highly sex dependent (95% adaptation in females and males after 114.9 ± 81.1 vs 65.4 ± 64.3 seconds, respectively, p < 0.00001). P wave duration was even less rate dependent than the PQ interval (9 ± 10% of linear QT/RR slopes). Rate corrected P wave duration was marginally but significantly shorter in females than in males (106.8 ± 8.4 vs 110.2 ± 7.9 ms, p < 0.00001). In addition to establishing physiologic standards, the study suggests that the curvatures and adaptation delay of the PQ/cycle-length dependency should be included in future non-invasive studies of atrial depolarizations.

Implications of Individual QT/RR Profiles-Part 1: Inaccuracies and Problems of Population-Specific QT/Heart Rate Corrections

Introduction

Universal QT correction formulas are potentially problematic in corrected QT (QTc) interval comparisons at different heart rates. Instead of individual-specific corrections, population-specific corrections are occasionally used based on QT/RR data pooled from all study subjects.

Objective

To investigate the performance of individual-specific and population-specific corrections, a statistical modeling study was performed using QT/RR data of 523 healthy subjects.

Methods

In each subject, full drug-free QT/RR profiles were available, characterized using non-linear regression models. In each subject, 50 baseline QT/RR readings represented baseline data of standard QT studies. Using these data, linear and log-linear heart rate corrections were optimized for each subject and for different groups of ten and 50 subjects. These corrections were applied in random combinations of heart rate changes between − 10 and + 25 beats per minute (bpm) and known QTc interval changes between − 25 and + 25 ms.

Results

Both the subject-specific and population-specific corrections based on the 50 baseline QT/RR readings tended to underestimate/overestimate the QTc interval changes when heart rate was increasing/decreasing, respectively. The result spread was much wider with population-specific corrections, making the estimates of QTc interval changes practically unpredictable.

Conclusion

Subject-specific heart rate corrections based on limited baseline drug-free data may lead to inconsistent results and, in the presence of underlying heart rate changes, may potentially underestimate or overestimate QTc interval changes. The population-specific corrections lead to results that are much more influenced by the combination of individual QT/RR patterns than by the actual QTc interval changes. Subject-specific heart rate corrections based on full profiles derived from drug-free baseline recordings with wide QT/RR distribution should be used when studying drugs expected to cause heart rate changes.

Electronic supplementary material

The online version of this article (10.1007/s40264-018-0736-1) contains supplementary material, which is available to authorized users.

Implications of Individual QT/RR Profiles-Part 2: Zero QTc/RR Correlations Do Not Prove QTc Correction Accuracy in Studies of QTc Changes

Introduction

In studies of drug-induced corrected QT (QTc) changes, fixed universal heart rate (HR) corrections (e.g., the Fridericia correction) are potentially misleading when assessing the effects of drugs that change HR. When data-specific corrections are designed, tests of their validity are needed. The proposed tests include zero correlations between QTc and corresponding RR values in the complete study data (pooling on-treatment and off-treatment interval measurements).

Objective

To document that this approach is potentially highly misleading, a statistical modeling study was conducted based on the full profiles of QT/RR data of 523 healthy subjects—254 females, mean age 33.5 years.

Methods

In each of the subjects, 50 baseline QT/RR readings were selected to model baseline data. In repeated experiments, groups of ten and 50 subjects were randomly selected and drug-induced HR increases between 0 and 25 beats per minute combined with QTc changes between − 20 and + 20 ms were modeled. In each experiment, subject-specific as well as population-specific HR corrections were designed so that the QTc interval data were uncorrelated to the corresponding RR interval data.

Results

The simulation experiments showed that when zero correlations of QTc data with RR data are combined with more than trivial HR increases, the HR corrections are substantially biased and underestimate or fully eliminate any drug-induced QTc interval changes. This result is in full agreement with theoretical considerations of HR correction principles.

Conclusions

The lack of correlation of QTc versus RR durations including on-treatment data does not prove any validity of HR corrections. Correlations of QTc versus RR in study data pooling on- and off-drug measurements should not be used to prove the appropriateness of HR corrections.

Electronic supplementary material

The online version of this article (10.1007/s40264-018-0735-2) contains supplementary material, which is available to authorized users.

Sex and race differences in J-Tend, J-Tpeak, and Tpeak-Tend intervals

To facilitate the precision of clinical electrocardiographic studies of J-to-Tpeak (JTp) and Tpeak-to-Tend (Tpe) intervals, the study investigated their differences between healthy females and males, and between subjects of African and Caucasian origin. In 523 healthy subjects (254 females; 236 subjects of African origin), repeated Holter recordings were used to measure QT, JT, JTp, and Tpe intervals preceded by both stable and variable heart rates. Subject-specific curvilinear regression models were used to obtain individual QTc, JTc, JTpc and Tpec intervals. Rate hysteresis, i.e., the speed with which the intervals adapted after heart rate changes, was also investigated. In all sex-race groups, Tpe intervals were not systematically heart rate dependent. Similar to QTc intervals, women had JTc, and JTpc intervals longer than males (difference 20–30 ms, p < 0.001). However, women had Tpec intervals (and rate uncorrected Tpe intervals) shorter by approximately 10 ms compared to males (p < 0.001). Subjects of African origin had significantly shorter QTc intervals than Caucasians (p < 0.001). Gradually diminishing race-difference was found for JTc, JTpc and Tpec intervals. JTc and JTpc were moderately increasing with age but Tpe/Tpec were not. Rate hysteresis of JTp was approximately 10% longer compared to that of JT (p < 0.001). In future clinical studies, Tpe interval should not be systematically corrected for heart rate and similar to the QT interval, the differences in JT, JTp and Tpe intervals should be corrected for sex. The differences in QT and JT, and JTp intervals should also be corrected for race.

Sources of QTc variability: Implications for effective ECG monitoring in clinical practice

Pharmaceuticals that prolong ventricular repolarization may be proarrhythmic in susceptible patients. While this fact is well recognized, schemes for sequential QTc interval monitoring in patients receiving QT‐prolonging drugs are frequently overlooked or, if implemented, underutilized in clinical practice. There are several reasons for this gap in day‐to‐day clinical practice. One of these is the perception that serially measured QTc intervals are subject to substantial variability that hampers the distinction between potential proarrhythmic signs and other sources of QTc variability. This review shows that substantial part of the QTc variability can be avoided if more accurate methodology for electrocardiogram collection, measurement, and interpretation is used. Four aspects of such a methodology are discussed. First, advanced methods for QT interval measurement are proposed including suggestion of multilead measurements in problematic recordings such as those in atrial fibrillation patients. Second, serial comparisons of T‐wave morphologies are advocated instead of simple acceptance of historical QTc measurements. Third, the necessity of understanding the pitfalls of heart rate correction is stressed including the necessity of avoiding the Bazett correction in cases of using QTc values for clinical decisions. Finally, the frequently overlooked problem of QT‐heart rate hysteresis is discussed including the possibility of gross QTc errors when correcting the QT interval for simultaneously measured short‐term heart rate.

Heart Rate Correction of the J-to-Tpeak Interval

Drug-induced changes of the J to T peak (JTp) and J to the median of area under the T wave (JT50) were reported to differentiate QT prolonging drugs that are predominant blockers of the delayed potassium rectifier current from those with multiple ion channel effects. Studies of drug-induced JTp/JT50 interval changes might therefore facilitate cardiac safety evaluation of new pharmaceuticals. It is not known whether formulas for QT heart rate correction are applicable to JTp and JT50 intervals. QT/RR, JTp/RR, and JT50/RR profiles were studied in 523 healthy subjects aged 33.5 ± 8.4 years (254 females). In individual subjects, 1,256 ± 220 electrocardiographic measurements of QT, JTp, and JT50 intervals were available including a 5-minute history of RR intervals preceding each measurement. Curvilinear, linear and log-linear regression models were used to characterize individual QT/RR, JTp/RR, and JT50/RR profiles both without and with correction for heart rate hysteresis. JTp/RR and JT50/RR hysteresis correction needs to be included but the generic universal correction for QT/RR hysteresis is also applicable to JTp/RR and JT50/RR profiles. Once this is incorporated, median regression coefficients of the investigated population suggest linear correction formulas JTpc = JTp + 0.150(1-RR) and JT50c = JT50 + 0.117(1-RR) where RR intervals of the underlying heart rate are hysteresis-corrected, and all measurements expressed in seconds. The established correction formulas can be proposed for future clinical pharmacology studies that show drug-induced heart rate changes of up to approximately 10 beats per minute.

Individually Rate Corrected QTc Intervals in Children and Adolescents

Accurate evaluation of the appearance of QTc sex differences during childhood and adolescence is intricate. Inter-subject differences of individual QT/RR patterns make generic heart rate corrections inaccurate because of fast resting heart rates in children. The study investigated 527 healthy children and adolescents aged 4–19 years (268 females, 50.9%). All underwent continuous ECG 12-lead monitoring while performing postural changes during a 70-min investigative protocol to obtain QT interval measurements at different heart rates. On average, more than 1200 ECG measurements (QT interval and its 5-min history of preceding RR intervals) were made in each subject. Curvilinear QT/RR regression involving intra-individual correction for QT/RR hysteresis were calculated in each subject. The projection of the QT/RR regressions to the heart rate of 60 beats per minute defined individually corrected QTc intervals. In males, gradual QTc shortening by about 15 ms appeared during the ages of 13–19 years synchronously with the incidence of secondary sex signs (p = 0.016). On the contrary, whilst gradual QTc prolongation by about 10 ms appeared in females, it occurred only during ages 16–19 years and was not related to the incidence of secondary sex signs (p = 0.18). The study also showed that in children and adolescents, linear QT/RR models fit the intra-subject data significantly more closely than the log-linear models (p < 0.001). The study speculates that hormonal shifts during puberty might be directly responsible for the QTc shortening in males but that QTc prolongation in females is likely more complex since it was noted to follow the appearance of secondary sex signs only after a considerable delay.

Detection of T Wave Peak for Serial Comparisons of JTp Interval

Electrocardiogram (ECG) studies of drug-induced prolongation of the interval between the J point and the peak of the T wave (JTp interval) distinguished QT prolonging drugs that predominantly block the delayed potassium rectifier current from those affecting multiple cardiac repolarisation ion channel currents. Since the peak of the T wave depends on ECG lead, a “global” T peak requires to combine ECG leads into one-dimensional signal in which the T wave peak can be measured. This study aimed at finding the optimum one-dimensional representation of 12-lead ECGs for the most stable JTp measurements. Seven different one-dimensional representations were investigated including the vector magnitude of the orthogonal XYZ transformation, root mean square of all 12 ECG leads, and the vector magnitude of the 3 dominant orthogonal leads derived by singular value decomposition. All representations were applied to the median waveforms of 660,657 separate 10-s 12-lead ECGs taken from repeated day-time Holter recordings in 523 healthy subjects aged 33.5 ± 8.4 years (254 women). The JTp measurements were compared with the QT intervals and with the intervals between the J point and the median point of the area under the T wave one-dimensional representation (JT50 intervals) by means of calculating the residuals of the subject-specific curvilinear regression models relating the measured interval to the hysteresis-corrected RR interval of the underlying heart rate. The residuals of the regression models (equal to the intra-subject standard deviations of individually heart rate corrected intervals) expressed intra-subject stability of interval measurements. For both the JTp intervals and the JT50 intervals, the curvilinear regression residuals of measurements derived from the orthogonal XYZ representation were marginally but statistically significantly lower compared to the other representations. Using the XYZ representation, the residuals of the QT/RR, JTp/RR and JT50/RR regressions were 5.6 ± 1.1 ms, 7.2 ± 2.2 ms, and 4.9 ± 1.2 ms, respectively (all statistically significantly different; p < 0.0001). The study concludes that the orthogonal XYZ ECG representation might be proposed for future investigations of JTp and JT50 intervals. If the ability of classifying QT prolonging drugs is further confirmed for the JT50 interval, it might be appropriate to replace the JTp interval since with JT50 it appears more stable.

Errors of Fixed QT Heart Rate Corrections Used in the Assessment of Drug-Induced QTc Changes

The accuracy of studies of drug-induced QTc changes depends, among others, on the accuracy of heart rate correction of QT interval. It has been recognized that when a drug leads to substantial heart rate changes, fixed universal corrections cannot be used and that alternative methods such as subject-specific corrections established for each study participant need to be considered. Nevertheless, the maximum heart rate change that permits use of fixed correction with reasonable accuracy has not been systematically investigated. We have therefore used full QT/heart-rate profiles of 751 healthy subjects (mean age 34.2 ± 9.6, range 18–61 years, 335 females) and compared their subject-specific corrections with 6 fixed corrections, namely Bazett, Fridericia, Framingham, Hodges, Rautaharju, and Sarma formulae. The comparison was based on statistical modeling experiments which simulated clinical studies of N = 10 or N = 50 female or male subjects. The experiments compared errors of ΔQTc intervals calculated as differences between QTc intervals at an initial heart rate (in the range of 40 to 120 beats per minute, bpm) and after a heart rate change (in the range from −20 to +20 bpm). The experiments also investigated errors due to spontaneous heart rate fluctuation and due to omission of correction for QT/RR hysteresis. In each experiment, the absolute value of the single-sided 90th percentile most remote from zero was used as the error estimate. Each experiment was repeated 10,000 times with random selection of modeled study group. From these repetitions, median and upper 80th percentile was derived and graphically displayed for all different combinations of initial heart rate and heart rate change. The results showed that Fridericia formula might be reasonable (with estimated errors of ΔQTc below 8 ms) in large studies if the heart rate does not change more than ± 10 bpm and that the errors by fixed corrections and the errors due to omission of QR/RR hysteresis are additive. Additionally, the results suggest that the variability introduced into QTc data by not correcting for the underlying heart rate accurately might have a greater impact in smaller studies. The errors by Framingham formula were practically the same as with the Fridericia formula. Other investigated fixed heart rate corrections led to larger ΔQTc errors.

Individual-Specific QT Interval Correction for Drugs With Substantial Heart Rate Effect Using Holter ECGs Extracted Over a Wide Range of Heart Rates

Although fixed QT correction methods are typically used to adjust for the effect of heart rate on the QT interval in thorough QT/QTc studies, individual-specific QT correction (QTcI = QT/RR^I ) is advisable for drugs that increase the heart rate by >5 to 10 beats/minute (bpm). QTcI is traditionally derived using resting drug-free electrocardiograms (ECGs) collected at prespecified times. However, the resting heart rate range in healthy individuals is narrow, and extrapolation of inferences from these data to higher heart rates could be inappropriate. Accordingly, the QTcI derived from triplicate ECGs extracted at prespecified times (the traditional [T] method, yielding QTcIT) was compared with QTcIs obtained using ECGs with a wider heart rate range (alternative Holter [H] method, yielding QTcIH) from 24-hour Holter recordings from 40 healthy individuals selected from a central ECG laboratory database. For QTcIH, 10-second ECGs were extracted at stable heart rates in the ranges of 51-60, 61-70, 71-80, and 81-90 bpm (9 ECGs in each bin = 36 ECGs). An independent set of 40 ECGs with heart rates from 51 to 90 bpm was extracted from each individual to validate the accuracy of QTcI by the 2 methods. For the validation set, the QTcIH was a better QT correction method (slope of QTc vs heart rate closer to zero) than QTcIT. The mean difference between QTcIT and QTcIH increased from 3.1 milliseconds at 65 bpm to 10.0 milliseconds at 90 bpm (P < 0.01). The QTcIT exceeded QTcIH at heart rates > 60 bpm. Employment of the QTcIH may be more appropriate for studies involving drugs that increase heart rate.

Importance of QT/RR hysteresis correction in studies of drug-induced QTc interval changes

QT/RR hysteresis and QT/RR adaptation are interlinked but separate physiological processes signifying how quickly and how much QT interval changes when heart rate changes, respectively. While QT interval duration is, as a rule, corrected for heart rate in terms of the QT/RR adaptation, the correction for QT/RR hysteresis is frequently omitted in studies of drug-induced QTc changes. This study used data from previously conducted thorough QT studies to investigate the extent of QTc errors caused by omitting the correction for QT/RR hysteresis, particularly in small clinical investigations. Statistical modeling approach was used to generate 11,000 simulated samples of 10-subject studies in which mixed effect PK/PD models were used to estimate drug-induced QTc changes at mean maximum plasma concentration of investigated compounds. Calculations of QTc intervals involving and omitting QT/RR hysteresis correction were compared. These comparisons showed that ignoring QT/RR hysteresis has two undesirable effects: (A) In the design of subject-specific heart rate corrections (needed in studies of drugs that change heart rate) omission of QT/RR hysteresis may lead to signals of QTc prolongation of more than 10 ms to be missed. (B) Irrespective of whether the investigated drug changes heart rate, omission of QT/RR hysteresis causes the widths of the confidence intervals of the PK/PD predicted QTc interval changes to be increased by 20–30% on average (exceeding 50% in some cases). This may lead to a failure of excluding meaningful QTc prolongation which would be excluded if using hysteresis correction. The study concludes that correction for QT/RR hysteresis should be incorporated into future studies of drug-induced QTc changes. Subject-specific heart rate corrections that omit hysteresis correction may lead to erroneously biased conclusions. Even when using universal (e.g. Fridericia) heart rate correction, hysteresis correction decreases the confidence intervals of QTc changes and thus helps avoiding false positive outcomes.

Theoretical and experimental comparison of lag-based and time-based exponential moving average models of QT hysteresis

OBJECTIVE

In the electrocardiogram, adaptation of the QT interval to variations in heart rate is not instantaneous. Quantification of this hysteresis phenomenon relies on mathematical models describing the relation between the RR and QT time series. These models reproduce hysteresis through an effective RR interval computed as a linear combination of the history of past RR intervals. This filter depends on a time constant parameter that may be used as a biomarker.

APPROACH

The most common hysteresis model is based on an autoregressive filter with an impulse response that decreases exponentially with the beat number (lag-based model). Recognizing that the QT time series is unevenly spaced, we propose two exponential moving average filters (time-based models) to define the effective RR interval: one with an impulse response that decreases exponentially with time in seconds, and one with a step response that relaxes exponentially with time in seconds. These two filters are neither linear nor time-invariant. Recurrence formulas are derived to enable efficient implementation.

MAIN RESULTS

Application to clinical signals recorded during tilt table test, exercise and 24 h Holter demonstrates that the three models perform similarly in terms of goodness-of-fit. When comparing the hysteresis time constant in two conditions with different heart rates, however, the time-based models are shown to reduce the bias on the hysteresis time constant caused by heart rate acceleration and deceleration.

SIGNIFICANCE

Time-based models should be considered when intergroup differences in both heart rate and QT hysteresis are expected.

Categorization and theoretical comparison of quantitative methods for assessing QT/RR hysteresis

BACKGROUND

In the human electrocardiogram, there is a lag of adaptation of the QT interval to heart rate changes, usually termed QT/RR hysteresis (QT-hys). Subject-specific quantifiers of QT-hys have been proposed as potential biomarkers, but there is no consensus on the choice of the quantifier.

METHODS

A comprehensive literature search was conducted to identify original articles reporting quantifiers of repolarization hysteresis from the surface ECG in humans.

RESULTS

Sixty articles fulfilled our inclusion criteria. Reported biomarkers were grouped under four categories. A simple mathematical model of QT/RR loop was used to illustrate differences between the methods. Category I quantifiers use direct measurement of QT time course of adaptation. They are limited to conditions where RR intervals are under strict control. Category IIa and IIb quantifiers compare QT responses during consecutive heart rate acceleration and deceleration. They are relevant when a QT/RR loop is observed, typically during exercise and recovery, but are not robust to protocol variations. Category III quantifiers evaluate the optimum RR memory in dynamic QT/RR relationship modeling. They estimate an intrinsic memory parameter independent from the nature of RR changes, but their reliability remains to be confirmed when multiple memory parameters are estimated. Promising approaches include the differentiation of short-term and long-term memory and adaptive estimation of memory parameters.

CONCLUSION

Model-based approaches to QT-hys assessment appear to be the most versatile, as they allow separate quantification of QT/RR dependency and QT-hys, and can be applied to a wide range of experimental settings.

Estimation of the QT-RR relation: trade-off between goodness-of-fit and extrapolation accuracy

Correction of the QT interval in the ECG for changes in heart rate (RR interval) is needed to compare groups of patients and assess the risk of sudden cardiac death. The QTc represents the QT interval at 60 bpm, although most patients typically have a faster heart rate, thus requiring extrapolation of the QT-RR relationship.

OBJECTIVE

This paper investigates the ability of QT-RR models with increasing number of parameters to fit beat-to-beat variations in the QT interval and provide a reliable estimate of the QTc.

APPROACH

One-, two- and three-parameter functions generalising the Bazett and Fridericia formulas were used in combination with hysteresis reduction (memory) obtained by time-averaging the history of RR intervals with exponentially-decaying weights. In normal men and women datasets of Holter recordings in normal subjects (24 h monitoring), two measures were computed for each model: the root mean square error (RMSE) of fitting and the difference between the estimated QTc and a reference QTc obtained by collecting data points around RR  =  1000 ms.

MAIN RESULTS

The two- and three-parameter functions all gave similar low RMSE with uncorrelated residues. An optimal memory parameter was found that still minimized the RMSE and could be used for all functions and subjects. This reduction in RMSE resulted from changes in the parameters linked to the increased steepness of the QT-RR relation after hysteresis reduction. At optimal memory, the two and three-parameter models provided poorer prediction of the QTc as compared to the Fridericia's model in subjects with fast heart rates, since accurate representation of the steeper QT-RR relation worsened the extrapolation that was then needed to determine the QTc.

SIGNIFICANCE

As a result, among all models investigated, the Fridericia formulation offered the best trade-off for QTc prediction robust to memory and fast heart rates.