The Effects of the Anesthetic Drug Propofol on the Way the Heart Recovers Between Beats

1. Purpose: to examine in detail the effect of propofol on Tp-e (an ECG measure of dispersion of repolarization); to search with maximum statistical power for a difference in this parameter before and after exposure to this widely used anaesthetic. To investigate the nature of any dose-response relationship between propofol and mean QTc and Tp-e intervals.
2. Hypotheses: 1. H0: mean pre-operative Tp-e = mean intra-operative Tp-e within each group vs. H1: mean pre-operative Tp-e ≠ mean intra-operative Tp-e within each group. 2. H0: mean intra-operative Tp-e group 1 = mean intra-operative Tp-e group 2 = mean intra-operative Tp-e group 3 vs. H1: mean intra-operative Tp-e group 1 ≠ mean intra-operative Tp-e group 2 ≠ mean intra-operative Tp-e group 3.
3. Justification: Propofol is an anesthetic agent that is widely used for induction and maintenance of anesthesia in children. It has long been thought that prolongation of repolarization, however caused, predisposes to a rare malignant ventricular tachyarrhythmia called torsades de pointes (TdP). The classic model for this hypothesis is a group of hereditary conditions collectively known as long QT syndrome. Although rare, this condition usually presents in childhood or early adulthood, with syncope, aborted cardiac arrest or sudden death, secondary to episodes of TdP. The genetic mutation affects the structure and function of myocardial potassium channels involved in repolarization dynamics. Some anaesthetic agents block some of these potassium channels, thus prolonging repolarization, producing an acquired long QT syndrome. QT interval prolongation per se is associated with, but is not the cause of, TdP. It has been shown recently that exaggeration of a physiological phenomenon called dispersion of repolarization (TDR) provides the right environment and the trigger for TdP. Normal TDR reflects the way that different layers of the myocardial wall repolarize at different rates - the outside fastest, then the inside \& finally the middle. Physiological TDR also determines the morphology of the T wave on the surface ECG. The interval between the peak and the end of the T wave is a measure of TDR. We therefore now have a new tool for assessing the risk posed by a drug that prolongs the QT interval. Evidence is accumulating that, if TDR is not increased, the risk of TdP is not increased, even if the QT interval is prolonged. Conversely, if TDR is exaggerated, the risk of TdP is raised, even if the absolute QT interval is within normal limits. In a pilot study, Whyte \& colleagues showed that propofol does not increase TDR, suggesting that the risk of TdP is not increased with this agent. That study examined only one dose at the extreme lower end of the range for surgical anesthesia \& had only 80% power. This study is designed to address those weaknesses and investigate more thoroughly the relationship between propofol and TDR, with the aim of being able to provide evidence-based recommendations, where none currently exist, on its use in patients with or at risk of long QT syndromes.
4. Objectives: a) to determine whether there is a significant difference between pre and post-induction mean QTc interval and mean Tp-e interval for each effect-site target concentration of propofol. b) to determine whether there is a relationship between propofol dose, and mean QTc and Tp-e intervals. The primary outcome of the study will be the presence or absence of differences in Tp-e within and between groups of children allocated by randomization to receive one of three therapeutic, clinically relevant, effect-site target concentrations of propofol. For each child, the endpoint of the study will be 5 minutes after induction of anesthesia.
5. Research Method: randomised, double-blinded within- and between groups comparative study in 60 unpremedicated ASA I-II children, aged between 3 and 10 years, undergoing procedural general anaesthesia. After obtaining written informed parental consent, and patient assent where appropriate, enrolled patients will be randomized to one of three groups, to receive a different steady state effect-site concentration of propofol. Block randomization will be prepared using computer generated random numbers. Allocation will be concealed using sealed sequentially numbered opaque envelopes. Prior to induction of anaesthesia, ECG electrodes will be sited at standardised locations for acquisition of a pre-operative 12 lead ECG. An intra-operative ECG, using the same electrode positions, will be taken 5 min after induction of anaesthesia, when the appropriate steady state effect-site concentration of propofol has been reached. The patient's involvement in the study will then be complete and the conduct of anaesthesia continued at the discretion of the supervising anaesthetist. All ECGs will be recorded in duplicate, at a paper speed of 50 mm/sec and with no identifying data or automated analysis on the recorded traces. Each ECG will be given a random number three-figure code, to allow identification of paired pre- and intra-operative traces after analysis. IV access will be obtained immediately before induction. Anaesthesia will be induced and maintained with propofol delivered by a syringe pump. After 5 minutes a steady state will have been reached at an effect-site concentration value of 3mcg/ml (group 1), 4.5mcg/ml (group 2) or 6mcg/ml (group 3). Throughout the study period, all children will receive routine monitoring. In an attempt to minimize sympathetic stimulation, laryngoscopy will not be permitted during the study period, and the airway will be maintained either by facemask or laryngeal mask. All the ECG traces will be analysed independently by two of the authors (SS and SW) in accordance with predetermined criteria. Both will be blinded to the study group and to the status of the ECG recording (pre- or intra-operative). Neither will be involved in recruitment or randomisation of patients, or in the conduct of the anaesthesia or acquisition of ECG recordings, all of which will be performed by one of the other investigators.

Data analysis: the QT and Tp-e intervals will be measured for all complete P-QRS-T cycles in leads II and V5 and averaged to give a mean QT interval and Tp-e interval for that lead. The QT interval will be measured from the start of the QRS complex to the end of the T-wave, defined as the point of return to the T-P baseline. If U waves are present, the end of the T-wave will be taken as the nadir of the curve between the T and U waves. The Tp-e interval will be measured from the peak of the T-wave to the end of the T-wave. Monophasic T wave peaks can be identified visually. For more complex T wave morphologies, the peak will be identified according to the criteria of Emori \& Antzelevitch.

Bland -Altman plots will be used to compare the ECG data from the two independent reviewers. Where an inter-observer difference of >10 msec in an RR interval or >20 msec in a QT or Tp-e interval is found, the recordings, still coded, will be re-analysed and a consensus reached if possible. Thus for each lead in each trace, two values for the mean RR interval, the mean QTc interval and the mean Tp-e interval, one from each independent reviewer, will be obtained. Each pair of values will then be averaged to give an overall value, which will then be used for further statistical analysis. Within-group and between-group comparisons of pre- and intra-operative ECG indices will be performed using two-way analysis of variance. Data analysis will be conducted by AC and SDW using Analyse-It® (Analyse-It software, Leeds, UK).

Sample size calculation: We have based our power calculations on results from a previous study carried out by Whyte et al. They found a mean (SD) Tp-e of 72.2 (10.9) msec in 49 pre-operative ECG traces from healthy children. The smallest ECG difference we can reliably detect is half of one small ECG square. At a paper speed of 50 mm/sec, this equates to 10 msec. However, the standard deviation Tp-e in pre-operative traces is 11 msec, so 10 ms is unlikely to be a clinically significant difference. In neonates on cisapride, Tp-e increased by a mean of 35 msec. Lubinski et al reported a mean increase in Tp-e of 17.2 msec in known (therefore presumably treated) LQTS adult patients. Searching for a larger difference would reduce the likelihood of any true difference being due to inter-observer variability. A sample size of 14 per group will detect a difference of 25 msec in Tp-e between the intra-operative means of the three groups with a power of 99% and the criterion for significance set at 0.003 ( 0.01 before a priori Bonferroni correction). In order to provide a small buffer in group sizes, to allow for unplanned exclusions, we plan to recruit 60 patients in total; 20 in each of the three groups