- Email: [email protected]
Identification of appropriate QTc formula in beagle dogs for nonclinical safety assessment
Regulatory Toxicology and Pharmacology 89 (2017) 118e124
Contents lists available at ScienceDirect
Regulatory Toxicology and Pharmacology journal homepage: www.elsevier.com/locate/yrtph
Identification of appropriate QTc formula in beagle dogs for nonclinical safety assessment Shitalkumar Patel, Laxit Bhatt*, Rajesh Patel, Chitrang Shah, Vipul Patel, Jitendra Patel, Rajesh Sundar, Upendra Bhatnagar, Mukul Jain Department of Pharmacology & Toxicology, Zydus Research Centre, Sarkhej-Bavla N.H. No. 8A, Moraiya, Ahmedabad, 382213 Gujarat, India
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 March 2017 Received in revised form 21 July 2017 Accepted 23 July 2017 Available online 25 July 2017
A number of drugs belonging to different therapeutic classes cause increase in QT interval duration, and this change has been associated with ventricular arrhythmias. Investigation of changes in QT intervals in toxicity studies in dogs is therefore of potential value. Estimation of a direct effect of drugs on the duration of the QT interval can be confused by drug-induced increases in heart rate. The objective of this evaluation was to identify an appropriate correction formula by comparing different formulae that could appropriately correct changes in QT interval in conscious beagle dogs in toxicology studies. Most commonly used QTc (QT correction) formulae are derived from human observations, like Bazett's formula and thus are not applicable for other species like dogs, where the normal values of heart rate is higher compared to humans. Using our historical data, we have established and compared different correction formulas and found that Van de Water's formula is the most appropriate for dog under conditions stated. However, there is no universally accepted formula for QTc calculation in dogs, and hence each organization should have its own formula, based on the analysis of data obtained from the strain used in its own experimental conditions. © 2017 Elsevier Inc. All rights reserved.
Keywords: Beagle dogs Electrocardiogram Heart rate QT correction formula Preclinical assessment
1. Introduction The development of a pharmaceutical is a stepwise process involving an evaluation of both animal and human efficacy and safety information. The nonclinical safety assessment for marketing approval of a pharmaceutical usually includes pharmacological studies, general toxicity and other toxicity studies (ICH M3(R2), 2009). Traditionally, registration of pharmaceuticals for human use requires the inclusion of a non-rodent species in the safety assessment process (CPMP & CPVP, 1997). The dog is the most frequently used non-rodent species for toxicity studies and safety evaluation (Griffin et al., 2004). ECG is one of the most important parameter in non-rodents for
Abbreviations: AAALAC, Association for Assessment and Accreditation of Laboratory Animal Care; ECG, Electrocardiogram; GLP, Good Laboratory Practice; HR, Heart Rate; ICH, International Council on Harmonisation; NCEs, New Chemical Entities; SD, Standard Deviation; QTc(B), Bazzet's formula; QTc(F), Fridericia's formula; QTc(FM), Framingham (Sagie)'s formula; QTc(V), Van de Water's formula; QTc(H), Hodges' formula; msec, millisecond. * Corresponding author. E-mail address: [email protected] (L. Bhatt). http://dx.doi.org/10.1016/j.yrtph.2017.07.026 0273-2300/© 2017 Elsevier Inc. All rights reserved.
evaluation of any candidate drug during nonclinical safety studies. Data derived from these evaluations are required prior to the commencement of successive stages (Tattersall et al., 2006). Various scientific literatures provide evidences on the Dog's role as the most relevant species to human electrophysiology (Gralinski, 2003). Adverse effects like, Torsades de Pointes (TdP), a particular type of cardiac arrhythmia, are associated with different pharmaceutical drugs. It is, hence, important to have a comprehensive evaluation of such cardiac risks during drug development process (Hanton, 2007). Assessment of markers of pro-arrhythmic risk is conducted during preclinical studies in dogs by measuring the QT interval prolongation, caused due to increased cardiac action potential, considered to be an indicator of arrhythmic risk (Guth et al., 2004). The QT interval, measured in the ECG, comprises the length from the onset of the QRS complex to the end of the T wave in the cardiac cycle, and represents ventricular depolarization and repolarization (Tilley, 1981). Heart rate and QT interval have an inverse relationship. The measurement of the QT interval, hence, should be corrected to improve its diagnostic purpose, which requires a formula that corrects the value of QT interval for increase in heart rate (Harada et al., 2009). Presently, different correction formulae are used, which have
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
been derived for analyzing human data and some for analyzing anaesthetized animals (Oliveira et al., 2014). Regulatory guidances ICH S7B and E14 do not clearly instruct on which specific QTc method should be adopted for preclinical evaluation. Most commonly used QTc formulae are derived from human observations, like Bazett's formula and may not be applicable for other species; for example in dogs, the normal values of heart rate is higher compared to humans, hence, rendering it unusable (Holzgrefe et al., 2014; Oliveira et al., 2014). The objective of this analysis was therefore to identify an appropriate correction formula by comparing different formulae that could appropriately correct changes in QT interval in conscious beagle dogs in toxicological studies with respect to heart rate, age and sex. This work was the subject of a poster presentation at the 8th RBF International Symposium, 2017, Ahmedabad 2e4 February. 2. Methods This retrospective analysis included clinically healthy conscious 183 beagle dogs (91 male, 92 female; Age 15e88 months) from different repeat dose toxicity studies conducted at Zydus Research Centre, Ahmedabad, India. These studies were approved by Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), Govt. of India for animal experimentation. The test facility is accredited with GLP certificate from National GLP Compliance Monitoring Authority, Govt. of India for the conduct of toxicity studies and additionally accredited by AAALAC International for animal ethics. ECGs were recorded in conscious dogs restrained in a standing position. All the procedures were performed in accordance with the Standard Operating Procedures (SOPs) in place at the institution. Electrode gel was used to improve conduction. The dogs were allowed to settle and once a good, clear signal was obtained, the chart recorders were actioned. ECGs were recorded using CARDIOVIT AT-1(VET) Electrocardiograph Machine, Schiller AG, Switzerland. Traces were recorded at a speed of 25 mm/s and ECG sensitivity at 5 mm/mV. Prior to the recording of ECGs, calibration of the instrument was performed at different speed and sensitivity. All the ECG traces were automatically assessed by the instrument, except for QTc, which was calculated using the correction formulae. Twelve leads ECG were recorded; HR and ECG intervals were determined from lead II. All dogs included in the studies had normal baseline ECGs, and no cardiac conduction abnormalities were identified. Some common features like the wandering baselines were found in traces in the ECGs. Intervals PR, QRS, RR, QT and HR were automatically measured by the instrument and printed on the ECG paper. The data were obtained from pre-dosed animals (baseline animals) and no pharmacologically challenged or test-item treated animals were used for the current analysis. Also, data from one animal has been included only one single time in same analysis. Five published correction formulae were considered for analysis
Table 1 QT correction formulae considered for this analysis. Name
Abbreviation
Formulae
Bazett Fridericia Framingham (Sagie) Van de Water Hodges
QTc(B) QTc(F) QTc(FM) QTc(V) QTc(H)
QT*(RR)1/2 QT*(RR)1/3 QT þ 0.154*(1000 eRR) QT-0.087(RR-1000) QTþ0.00175*(HR-60)
119
(Table 1): Bazett; Fridericia; Sagie, Larson, Goldberg, Bengston and Levy; Van de Water, Verheyen, Xhonneux and Reneman and Hodges M, Salerno D, Erlien D. From these formulae, Van de Water, has been designed to correct QT intervals of anaesthetized dogs, while the others are primarily designed to correct human QT intervals. The QTc and RR intervals were then graphed on a scatter plot, with the QTc (milliseconds) on Y-axis and RR interval (milliseconds) on X-axis. Five different QTc-RR interval scatter plots were generated, one for each QTc formula (Fig. 2). The slope of QTc-RR regression line for each QTc formula was determined and used to compare QTc formulae. Regression line slopes close to zero indicate consistency in calculating QTc values across the range of heart rates. A linear aggression model containing terms for sex, age and HR measurement was used to assess the relationship between variables in this analysis. The upper and lower 95% confidence limits are also presented along with the p-value for significance. Animals were also stratified by age, sex, and heart rate, according to the following groups: Age: 15e35 months and >35 months; Heart rate: >120 and 120; and Sex: males and females. QTc- RR scatter plots, means, standard deviations, and ranges were evaluated for each group (Table 2). 3. Results Our data comprised of ECGs from 183 beagle dogs. All ECGs measurements were taken from lead II. QTc based on the five correction formulae for each of the sub-groups (heart rate, sex, age & all samples) and parameters (Mean, SD, 2 SD, range, slope & p value) for individual QTc correction formula is shown in Table 2. There was positive correlation between the uncorrected QT intervals and RR intervals, for all samples as shown in Fig. 1. The QTc-RR interval (all samples) scatter plots and regression lines based on the Bazett, Fridericia, Van de Water, Hodges, and Framingham formulas are shown in Fig. 2. Upon evaluation of all the data, Van de Water's formula gave a regression line with a slope closest to zero (0.001) indicating the best consistency. Regression lines for other formula were: Hodges (0.229), Bazett (0.110), Fridericia (0.032), and Framingham (0.066) with values on different distances from zero indicating different levels of their consistency. To consider the data according heart rate (Above and below 120), sex (male and female) and Age (15e35 and above 35 month), Bazett's formula gave a regression line with a slope closest to zero (0.039), indicating the best consistency across heart rates (>120) (Fig. 3); while the Van de Water's formula gave a regression line with a slope closest to zero (0.020), (0.005), (0.007), (0.018) indicating the best consistency respectively across heart rates (120), male, female and age (15e35) (Fig. 4). The Fridericia's formula gave a regression line with a slope closest to zero (0.014) indicating the best consistency in age (>35) (Fig. 5). While, Hodges's formula gave a regression line with a slope closest to zero indicating less consistency compared to other formulae in respect to age, sex and heart rate. To consider the 2 SD thresholds, Van de Water's formula gave the lowest values (284) across the all samples while Hodges's formula gave the highest values (351). We found similar results across the heart rate, age and sex (Table 2). Based on parameters like Range (Minimum to maximum), Van de Water's formula gave the narrowest range (155) of QTc values across the all samples while Bazett's formula gave a broad range (192) of QTc values. When divided into subgroups (age, sex, heart rate), Van de Water's formula gave the narrowest range of QTc values across heart rate (120), sex and age. The Hodges' formula gave the narrowest range of QTc values across heart rate (>120)
120
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
Table 2 Calculated QTc (msec) values with parameters for all sub-groups. Parameters
QTc Formula
Slope
Mean
SD
Min. e Max.
Range
2*SD
95% Confidence Limit
p- value
All samples
Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's
0.110 0.032 0.001 0.229 0.066 0.039 0.280 0.571 0.976 0.330 0.112 0.044 0.020 0.181 0.087 0.104 0.027 0.005 0.224 0.062 0.121 0.041 0.007 0.240 0.075 0.123 0.048 0.018 0.221 0.086 0.093 0.014 0.018 0.224 0.048
272 248 243 286 270 287 250 239 325 276 266 245 242 277 267 272 248 243 286 270 270 245 239 289 268 266 244 240 279 266 275 249 242 295 272
29 24 21 33 22 27 24 19 23 19 23 20 17 23 19 29 25 21 33 22 22 17 14 29 16 20 15 13 25 15 29 25 21 34 22
225e417 218e391 216e372 226e396 230e393 258e407 223e360 216e327 294e396 257e362 219e417 206e391 208e372 220e393 225e393 225e417 218e391 216e372 226e396 230e393 219e358 206e312 208e285 220e373 225e323 222e315 206e274 208e267 226e357 229e294 219e417 215e391 216e372 220e396 225e393
192 173 156 170 163 149 137 111 102 106 198 185 164 173 168 192 173 155 170 163 139 106 77 153 98 93 68 59 131 64 198 176 155 176 168
329 297 284 351 314 341 298 277 371 313 313 286 276 323 305 330 298 285 352 315 314 279 267 346 300 305 274 265 329 296 334 299 285 363 316
0.13930 0.05848 0.02250 0.25230 0.08917 0.21150 0.09009 0.00862 0.52300 0.07950 0.15140 0.07985 0.05208 0.21170 0.11880 0.15390 0.07178 0.03451 0.26290 0.10120 0.15660 0.07431 0.03512 0.26740 0.10180 0.15700 0.07910 0.04587 0.24820 0.11270 0.14220 0.05889 0.01982 0.26290 0.08616
<0.0001 <0.0295 0.9365 <0.0001 <0.0001 0.8634 0.2670 0.0633 <0.0001 0.3253 <0.0001 0.0212 0.2124 <0.0001 <0.0001 <0.0001 0.2560 0.8178 <0.0001 0.0020 <0.0001 0.0125 0.5682 <0.0001 <0.0001 <0.0001 0.0034 0.1774 <0.0001 <0.0001 0.0003 0.5491 0.3417 <0.0001 0.0137
Heart Rate >120 (n ¼ 41)
Heart Rate 120 (n ¼ 142)
Male (n ¼ 91)
Female (n ¼ 92)
Age 15e35 Months (n ¼ 88)
Age >35 Months (n ¼ 95)
to 0.07919 to 0.00330 to 0.02441 to-0.20480 to-0.04231 to 0.25110 to 0.31630 to 0.30430 to 0.20530 to 0.23370 to 0.07288 to 0.00689 to 0.01146 to 0.15020 to 0.05530 to 0.05488 to 0.01941 to 0.04357 to 0.18420 to 0.02333 to 0.08493 to 0.00915 to 0.01942 to 0.21170 to 0.04716 to 0.08900 to 0.01619 to 0.00864 to 0.19530 to 0.05824 to 0.04327 to 0.03155 to 0.05649 to 0.18480 to 0.01002
Key: SD: Standard Deviation; 2SD: 2SD threshold; CL: Confidence limit; Min.: Minimum; Max.: Maximum.
Fig. 1. Uncorrected QT and RR interval.
while, a broad range of QTc values was obtained amongst females and age (15e35 months). The Bazett's formula gave a broad range of QTc values across heart rates (120 and > 120), age (>35 months) and amongst males. In p-value evaluation for all samples analysis, we found that the correction formulae of Bazett, Fridericia, Hodges, and Framingham, the slopes of the QTc verses RR interval relationships were all highly statistically significantly different from zero and therefore were not considered to have adequately corrected for heart rate changes;
while, the correction formula that was considered to have most adequately normalized the QT interval was that of Van de Water's formula, p ¼ 0.9365 because it was not statistically significant different from zero. When the data was divided into subgroups, only one correction formula (Hodges's formula) was found to be highly statistically significant different from zero (p < 0.0001) across the heart rate (>120) and remaining four formulae were not statistically significant different from zero, hence it was considered to have adequately normalized the QT interval, but among them Bazett's formula (p value ¼ 0.8634) was considered to most adequately normalized the QT interval compared to other formulae. In case of sex (male), and age (>35), three formula (Bazett, Hodges, and Framingham formulae) were statistically significant different from zero and remaining two formulae (Van de water and Fridericia formulae) were not statistically significant different from zero, in which Van de Water's formula (p ¼ 0.8178) for male and Fridericia's formulae (p ¼ 0.5491) were considered to most adequately normalized the QT interval compared to other formulae in case of age (>35).
4. Discussion The International Conference on Harmonisation (ICH) recommends that cardiovascular safety pharmacological studies and repeat dose non-rodent toxicity studies be performed, most typically in the dog or primate for any new chemical entity (ICH M3 (R2), 2009). Recording the surface electrocardiogram has been the basic gold standard for determining effects of pharmaceutical compounds on cardiac electrophysiology. A high correlation is
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
121
Fig. 2. Corrected QT intervals for different formulae.
found between action potential prolongation in dog purkinje fibers and increasing QT interval in humans with positive control pharmaceuticals (Gintant et al., 2001). It is important to accurately evaluate the potential of all new candidate drugs to prolong the QT interval preclinical studies (Hammond et al., 2001). Studies have found a negative correlation between HR and QT interval (Spence et al., 1998). Such findings highlight the importance of evaluation of QT interval, by using a correction formula, as variations in HR may mask abnormalities in QT interval. It is difficult to evaluate the clinical and biological significance of minor QT changes, even when they are statistically significant because of the numerous sources of inaccuracies with regard to QT measurement (Moss, 1999). Differences in HR can be a consequence
not only of autonomic conditioning but also of an external factor, like a drug (Malik, 2004). It is very interesting that even during the highly variable HR of SA in the dog; the QT does not change due to QT “memory.” QT memory holds that QT is determined by the average HR measured from the three to six beats preceding the QT being measured (Hamlin et al., 2004). A number of correction formulae are currently used, most of which were derived for analyzing human data, and some for analyzing anaesthetized animals. However little information is available regarding the correction of QT for changes in HR in conscious restrained beagle dogs in toxicology environment (Hanton et al., 2001). For example, in a recent survey, 41% of the 74 laboratories that responded, were using Bazett's formula despite
122
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
Fig. 5. Fridericia's formula for age >35 months. Fig. 3. Bazett's formula for HR > 120.
Fig. 4. Van de water's formula and individual parameters.
the fact that this formula is generally recognized as being unsuitable for correcting dog QT intervals (Hanton et al., 2001; Haverkamp et al., 2000). Our analysis of in-house generated data supports the previous claims of unsuitability of Bazett's and Fridericia's formulae for correction of QT interval in conscious dogs. Moreover, most studies have been done in respect to increase or decrease heart rate means across the heart range against different QTc formula but in current studies, we have analyzed the appropriate QTc formula for dog by considering all pre dose data and also
evaluated after tracing the data w.r.t heart rate, age and sex (Table 2). Previous work comparing various QT correction formulas has produced some conflicting results. One study has concluded that Bazett's and Fridericia's correction formula should be avoided to correct QT interval duration for change in Heart rate in conscious dogs (Hammond et al., 2001; Hanton et al., 2001) while Van de Water's correction factor can be used in toxicology assessment to correct QT interval duration for increases in HR. Our data supports
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
123
Table 3 Basis for recommendations. Recommendations
Basis
Larger sample size
Larger sample size is more representative of a population. It limits the influence of extreme observations or outliers. A large sample size broadens the range of possible data and forms a better picture for analysis. Non-invasive telemetry system Non-invasive telemetry experiments offer long and continuous recordings and simultaneous average acquisitions. Other advantages being: freely moving animals, no surgery, and low stress levels. Institutions must have own QT correction formula To nullify the variations arising due to difference in methodologies regarding conducting the experiments like with its historical data instrument types (Automatic/manual interpretation), source of animals, laboratory conditions and animal husbandry practices. Positive control To assess the reliability of the correction formulae over a wide range of HR
this claim when all samples were considered but when we traced the data for increase in HR above 120 (beats/minute), Bazett's formula was found to be more appropriate than others. This was may be due to insufficient sample size of HR > 120. Hence, we suggest similar data analysis with higher sample size. Given these inconsistencies, some investigators have suggested that separate rate correction formulae may be appropriate for different HR ranges (Rautaharju et al., 2009). Using control data from beagle dogs used in the toxicology assessment, two correction factors described by Bazett (1992) and Van de Water (1989) were identified to adequately correct QT intervals for changes in HR (i.e. slope of QTc versus RR not significantly different from zero). Although both adequately corrected QT intervals, Van de Water's formula, however, showed a statistically superior correction and therefore was identified for use in the evaluation of compound effects. The influence of sex on the evaluation of QT correction was also assessed and showed that sex did not influence the outcome of the evaluations (Fig. 4, Table 2). However, age had a slight effect on the p-value of the formulae, where Fridericia's formula was more suitable than Bazett's formula (Table 2; Fig. 5). We found that Van de Water's formula provides the most consistent correction across a wide range of heart rates. As it can be inferred from Table 2, slope for Van de Water's formula was 0.001 (closest to zero), 2 SD and range of QTc values were 284 and 155 respectively, which are very less compare to other formulae used in this analysis. These values most definitely derive that Van de Water's formula is the most consistent and has lesser variation than other formulae routinely used. Our above conclusion is supported by p-value obtained by Van de Water's formula, which is 0.9635, highly non-significant compared to other formulae. Although, as described before, when HR was considered for >120 (beats/minute), Bazett's formula obtained p-value of 0.8634, making it more suitable to correct QT intervals. Additionally, this analysis was performed using digital electrocardiography with amplification of the initial ECG data to allow for more precise measurements than manual measurement. This instrument gives a direct value of HR, PR, RR, QRS, QT, and hence it is easier to calculate QTc with different formula with accurate results. A limitation inherent to QT measurement occurs when the T and U wave are superimposed or cannot be separated, or when the T and P waves are superimposed, resulting in an inadvertent manual error in measurements of different intervals and thus in QTc values. The analysis performed in this study was performed in accordance with current Recommendations for the Standardization and Interpretation of the ECG (Rautaharju et al., 2009). In conclusion, Van de Water's QTc formula can be used in toxicology assessment to correct QT interval duration across HR and in cases on increased HR; Bazett's formula also has an important role to play. Data is still need to be collected to cover a wide range of HR and further validation is required due to sample size limitation.
Additionally, the development and introduction of newer technologies such as non-invasive telemetry systems that allows continuous ECG recordings in freely moving dogs in a toxicology environment should significantly enhance the quality, reliability, power and ultimately the value of ECG assessment in dog toxicology studies. Organizations can also conduct studies with a positive control to assess the reliability of the correction formulae over a wide range of HR. The authors also recommend institutions to determine appropriate correction formula using its own Historical QT data from its particular strain of dogs and under determined experimental conditions (Table 3). Acknowledgements The authors would like to thank the Study Directors and Study Personnel involved in the designing, conducting and reporting the studies in current evaluation. All the toxicity studies included under this analysis were financed and performed at Zydus Research Centre, Cadila Healthcare Ltd., Ahmedabad, India. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.yrtph.2017.07.026. Transparency document Transparency document related to this article can be found online at http://dx.doi.org/10.1016/j.yrtph.2017.07.026. References Gintant, G. a, Limberis, J.T., McDermott, J.S., Wegner, C.D., Cox, B.F., 2001. The canine Purkinje fiber: an in vitro model system for acquired long QT syndrome and drug-induced arrhythmogenesis. J. Cardiovasc. Pharmacol. 37, 607e618. http:// dx.doi.org/10.1097/00005344-200105000-00012. Gralinski, M.R., 2003. The dog's role in the preclinical assessment of QT interval prolongation. Toxicol. Pathol. (31 Suppl), 11e16. http://dx.doi.org/10.1080/ 01926230390174887. Griffin, G., Stokes, W.S., Pakes, S.P., Gauthier, C., 2004. The ICLAS/CCAC international symposium on regulatory testing and animal welfare. ATLA Altern. Lab. Anim 32, 707e712. Guideline, Committee for Proprietary Medicinal Products (CPMP) & Committee for Veterinary Medicinal Products. Note for Guidance on Inclusion of Antioxidants and Antimicrobial Preservatives in Medicinal Products. 1997. Retreived from: http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_ guideline/2009/09/WC500003408.pdf Guideline, 2009. ICH Harmonised Tripartite. “Guidance on nonclinical safety studies for the conduct of human clinical trials and marketing authorization for pharmaceuticals M3 (R2).”. In: International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Retreived from: http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/ Guidelines/Multidisciplinary/M3_R2/Step4/M3_R2__Guideline.pdf. Guth, B.D., Germeyer, S., Kolb, W., Markert, M., 2004. Developing a strategy for the nonclinical assessment of proarrhythmic risk of pharmaceuticals due to prolonged ventricular repolarization. J. Pharmacol. Toxicol. Methods 49, 159e169. http://dx.doi.org/10.1016/j.vascn.2004.02.006. Hamlin, R.L., Kijtawornrat, A., Keene, B.W., 2004. How many cardiac cycles must be
124
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
measured to permit accurate RR, QT, and QTc estimates in conscious dogs? J. Pharmacol. Toxicol. Methods 50, 103e108. http://dx.doi.org/10.1016/ j.vascn.2004.03.013. Hammond, T.G., Carlsson, L., Davis, A.S., Lynch, W.G., MacKenzie, I., Redfern, W.S., Sullivan, A.T., Camm, A.J., 2001. Methods of collecting and evaluating nonclinical cardiac electrophysiology data in the pharmaceutical industry: results of an international survey. Cardiovasc. Res. 49, 741e750. http://dx.doi.org/ 10.1016/S0008-6363(00)00310-2. Hanton, G., 2007. Preclinical cardiac safety assessment of drugs. Drugs R. D. http:// dx.doi.org/10.2165/00126839-200708040-00002. Hanton, G., Nahas, K., Priou, C., Rabemampianina, Y., Baneux, P., 2001. The Qt interval in the dog Ecg: importance in preclinical toxicology and relationship to heart rate. Toxicol. Methods 11, 21e40. http://dx.doi.org/10.1080/ 105172301300055133. Harada, T., Ishizaki, F., Hamada, M., Horie, N., Nitta, Y., Nitta, K., Katsuoka, H., Nakamura, S., 2009. Circadian rhythm of heart-rate variability and autonomic cardiovascular regulation in Parkinson's disease. Auton. Neurosci. Basic Clin. 46, 333e340. http://dx.doi.org/10.1016/j.autneu.2010.07.016. Haverkamp, W., Breithardt, G., Camm, J. a, Janse, M.J., 2000. The potential for QT prolongation and proarrhythmia by non-antiarrhythmic drugs: clinical and regulatory implications. Eur. Hear. J. 47, 219e233. http://dx.doi.org/10.1053/ euhj.2000.2249. Holzgrefe, H., Ferber, G., Champeroux, P., Gill, M., Honda, M., Greiter-Wilke, A., Baird, T., Meyer, O., Saulnier, M., 2014. Preclinical QT safety assessment: cross-
species comparisons and human translation from an industry consortium. J. Pharmacol. Toxicol. Methods 69, 61e101. http://dx.doi.org/10.1016/ j.vascn.2013.05.004. Malik, M., 2004. Errors and misconceptions in ECG measurement used for the detection of drug induced QT interval prolongation. J. Electrocardiol. 37, 25e33. http://dx.doi.org/10.1016/j.jelectrocard.2004.08.005. Moss, A., 1999. The QT interval and torsade de pointes. Drug Saf. 21, 5e10. Oliveira, M.S., Muzzi, R.A.L., Muzzi, L.A.L., Cherem, M., Mantovani, M.M., 2014. QT interval in healthy dogs: which method of correcting the QT interval in dogs is appropriate for use in small animal clinics? Pesqui. Vet. Bras. 34, 469e472. http://dx.doi.org/10.1590/S0100-736X2014000500014. Rautaharju, P.M., Surawicz, B., Gettes, L.S., 2009. Aha/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram. Part IV: the ST segment, T and U Waves, and the QT interval a scientific statement from the american heart association electrocardiography and arrhythmias Co. J. Am. Coll. Cardiol. 53, 982e991. http://dx.doi.org/10.1016/j.jacc.2008.12.014. Spence, S., Soper, K., Hoe, C.M., Coleman, J., 1998. The heart rate-corrected QT interval of conscious beagle dogs: a formula based on analysis of covariance. Toxicol. Sci. 45, 247e258. http://dx.doi.org/10.1006/toxs.1998.2529. Tattersall, M.L., Dymond, M., Hammond, T., Valentin, J.P., 2006. Correction of QT values to allow for increases in heart rate in conscious Beagle dogs in toxicology assessment. J. Pharmacol. Toxicol. Methods 53, 11e19. http://dx.doi.org/10.1016/ j.vascn.2005.02.005. Tilley, L.P., 1981. Basic canine and feline electrocardiography. Can. Vet. J. 22, 23e25.
Contents lists available at ScienceDirect
Regulatory Toxicology and Pharmacology journal homepage: www.elsevier.com/locate/yrtph
Identification of appropriate QTc formula in beagle dogs for nonclinical safety assessment Shitalkumar Patel, Laxit Bhatt*, Rajesh Patel, Chitrang Shah, Vipul Patel, Jitendra Patel, Rajesh Sundar, Upendra Bhatnagar, Mukul Jain Department of Pharmacology & Toxicology, Zydus Research Centre, Sarkhej-Bavla N.H. No. 8A, Moraiya, Ahmedabad, 382213 Gujarat, India
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 March 2017 Received in revised form 21 July 2017 Accepted 23 July 2017 Available online 25 July 2017
A number of drugs belonging to different therapeutic classes cause increase in QT interval duration, and this change has been associated with ventricular arrhythmias. Investigation of changes in QT intervals in toxicity studies in dogs is therefore of potential value. Estimation of a direct effect of drugs on the duration of the QT interval can be confused by drug-induced increases in heart rate. The objective of this evaluation was to identify an appropriate correction formula by comparing different formulae that could appropriately correct changes in QT interval in conscious beagle dogs in toxicology studies. Most commonly used QTc (QT correction) formulae are derived from human observations, like Bazett's formula and thus are not applicable for other species like dogs, where the normal values of heart rate is higher compared to humans. Using our historical data, we have established and compared different correction formulas and found that Van de Water's formula is the most appropriate for dog under conditions stated. However, there is no universally accepted formula for QTc calculation in dogs, and hence each organization should have its own formula, based on the analysis of data obtained from the strain used in its own experimental conditions. © 2017 Elsevier Inc. All rights reserved.
Keywords: Beagle dogs Electrocardiogram Heart rate QT correction formula Preclinical assessment
1. Introduction The development of a pharmaceutical is a stepwise process involving an evaluation of both animal and human efficacy and safety information. The nonclinical safety assessment for marketing approval of a pharmaceutical usually includes pharmacological studies, general toxicity and other toxicity studies (ICH M3(R2), 2009). Traditionally, registration of pharmaceuticals for human use requires the inclusion of a non-rodent species in the safety assessment process (CPMP & CPVP, 1997). The dog is the most frequently used non-rodent species for toxicity studies and safety evaluation (Griffin et al., 2004). ECG is one of the most important parameter in non-rodents for
Abbreviations: AAALAC, Association for Assessment and Accreditation of Laboratory Animal Care; ECG, Electrocardiogram; GLP, Good Laboratory Practice; HR, Heart Rate; ICH, International Council on Harmonisation; NCEs, New Chemical Entities; SD, Standard Deviation; QTc(B), Bazzet's formula; QTc(F), Fridericia's formula; QTc(FM), Framingham (Sagie)'s formula; QTc(V), Van de Water's formula; QTc(H), Hodges' formula; msec, millisecond. * Corresponding author. E-mail address: [email protected] (L. Bhatt). http://dx.doi.org/10.1016/j.yrtph.2017.07.026 0273-2300/© 2017 Elsevier Inc. All rights reserved.
evaluation of any candidate drug during nonclinical safety studies. Data derived from these evaluations are required prior to the commencement of successive stages (Tattersall et al., 2006). Various scientific literatures provide evidences on the Dog's role as the most relevant species to human electrophysiology (Gralinski, 2003). Adverse effects like, Torsades de Pointes (TdP), a particular type of cardiac arrhythmia, are associated with different pharmaceutical drugs. It is, hence, important to have a comprehensive evaluation of such cardiac risks during drug development process (Hanton, 2007). Assessment of markers of pro-arrhythmic risk is conducted during preclinical studies in dogs by measuring the QT interval prolongation, caused due to increased cardiac action potential, considered to be an indicator of arrhythmic risk (Guth et al., 2004). The QT interval, measured in the ECG, comprises the length from the onset of the QRS complex to the end of the T wave in the cardiac cycle, and represents ventricular depolarization and repolarization (Tilley, 1981). Heart rate and QT interval have an inverse relationship. The measurement of the QT interval, hence, should be corrected to improve its diagnostic purpose, which requires a formula that corrects the value of QT interval for increase in heart rate (Harada et al., 2009). Presently, different correction formulae are used, which have
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
been derived for analyzing human data and some for analyzing anaesthetized animals (Oliveira et al., 2014). Regulatory guidances ICH S7B and E14 do not clearly instruct on which specific QTc method should be adopted for preclinical evaluation. Most commonly used QTc formulae are derived from human observations, like Bazett's formula and may not be applicable for other species; for example in dogs, the normal values of heart rate is higher compared to humans, hence, rendering it unusable (Holzgrefe et al., 2014; Oliveira et al., 2014). The objective of this analysis was therefore to identify an appropriate correction formula by comparing different formulae that could appropriately correct changes in QT interval in conscious beagle dogs in toxicological studies with respect to heart rate, age and sex. This work was the subject of a poster presentation at the 8th RBF International Symposium, 2017, Ahmedabad 2e4 February. 2. Methods This retrospective analysis included clinically healthy conscious 183 beagle dogs (91 male, 92 female; Age 15e88 months) from different repeat dose toxicity studies conducted at Zydus Research Centre, Ahmedabad, India. These studies were approved by Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), Govt. of India for animal experimentation. The test facility is accredited with GLP certificate from National GLP Compliance Monitoring Authority, Govt. of India for the conduct of toxicity studies and additionally accredited by AAALAC International for animal ethics. ECGs were recorded in conscious dogs restrained in a standing position. All the procedures were performed in accordance with the Standard Operating Procedures (SOPs) in place at the institution. Electrode gel was used to improve conduction. The dogs were allowed to settle and once a good, clear signal was obtained, the chart recorders were actioned. ECGs were recorded using CARDIOVIT AT-1(VET) Electrocardiograph Machine, Schiller AG, Switzerland. Traces were recorded at a speed of 25 mm/s and ECG sensitivity at 5 mm/mV. Prior to the recording of ECGs, calibration of the instrument was performed at different speed and sensitivity. All the ECG traces were automatically assessed by the instrument, except for QTc, which was calculated using the correction formulae. Twelve leads ECG were recorded; HR and ECG intervals were determined from lead II. All dogs included in the studies had normal baseline ECGs, and no cardiac conduction abnormalities were identified. Some common features like the wandering baselines were found in traces in the ECGs. Intervals PR, QRS, RR, QT and HR were automatically measured by the instrument and printed on the ECG paper. The data were obtained from pre-dosed animals (baseline animals) and no pharmacologically challenged or test-item treated animals were used for the current analysis. Also, data from one animal has been included only one single time in same analysis. Five published correction formulae were considered for analysis
Table 1 QT correction formulae considered for this analysis. Name
Abbreviation
Formulae
Bazett Fridericia Framingham (Sagie) Van de Water Hodges
QTc(B) QTc(F) QTc(FM) QTc(V) QTc(H)
QT*(RR)1/2 QT*(RR)1/3 QT þ 0.154*(1000 eRR) QT-0.087(RR-1000) QTþ0.00175*(HR-60)
119
(Table 1): Bazett; Fridericia; Sagie, Larson, Goldberg, Bengston and Levy; Van de Water, Verheyen, Xhonneux and Reneman and Hodges M, Salerno D, Erlien D. From these formulae, Van de Water, has been designed to correct QT intervals of anaesthetized dogs, while the others are primarily designed to correct human QT intervals. The QTc and RR intervals were then graphed on a scatter plot, with the QTc (milliseconds) on Y-axis and RR interval (milliseconds) on X-axis. Five different QTc-RR interval scatter plots were generated, one for each QTc formula (Fig. 2). The slope of QTc-RR regression line for each QTc formula was determined and used to compare QTc formulae. Regression line slopes close to zero indicate consistency in calculating QTc values across the range of heart rates. A linear aggression model containing terms for sex, age and HR measurement was used to assess the relationship between variables in this analysis. The upper and lower 95% confidence limits are also presented along with the p-value for significance. Animals were also stratified by age, sex, and heart rate, according to the following groups: Age: 15e35 months and >35 months; Heart rate: >120 and 120; and Sex: males and females. QTc- RR scatter plots, means, standard deviations, and ranges were evaluated for each group (Table 2). 3. Results Our data comprised of ECGs from 183 beagle dogs. All ECGs measurements were taken from lead II. QTc based on the five correction formulae for each of the sub-groups (heart rate, sex, age & all samples) and parameters (Mean, SD, 2 SD, range, slope & p value) for individual QTc correction formula is shown in Table 2. There was positive correlation between the uncorrected QT intervals and RR intervals, for all samples as shown in Fig. 1. The QTc-RR interval (all samples) scatter plots and regression lines based on the Bazett, Fridericia, Van de Water, Hodges, and Framingham formulas are shown in Fig. 2. Upon evaluation of all the data, Van de Water's formula gave a regression line with a slope closest to zero (0.001) indicating the best consistency. Regression lines for other formula were: Hodges (0.229), Bazett (0.110), Fridericia (0.032), and Framingham (0.066) with values on different distances from zero indicating different levels of their consistency. To consider the data according heart rate (Above and below 120), sex (male and female) and Age (15e35 and above 35 month), Bazett's formula gave a regression line with a slope closest to zero (0.039), indicating the best consistency across heart rates (>120) (Fig. 3); while the Van de Water's formula gave a regression line with a slope closest to zero (0.020), (0.005), (0.007), (0.018) indicating the best consistency respectively across heart rates (120), male, female and age (15e35) (Fig. 4). The Fridericia's formula gave a regression line with a slope closest to zero (0.014) indicating the best consistency in age (>35) (Fig. 5). While, Hodges's formula gave a regression line with a slope closest to zero indicating less consistency compared to other formulae in respect to age, sex and heart rate. To consider the 2 SD thresholds, Van de Water's formula gave the lowest values (284) across the all samples while Hodges's formula gave the highest values (351). We found similar results across the heart rate, age and sex (Table 2). Based on parameters like Range (Minimum to maximum), Van de Water's formula gave the narrowest range (155) of QTc values across the all samples while Bazett's formula gave a broad range (192) of QTc values. When divided into subgroups (age, sex, heart rate), Van de Water's formula gave the narrowest range of QTc values across heart rate (120), sex and age. The Hodges' formula gave the narrowest range of QTc values across heart rate (>120)
120
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
Table 2 Calculated QTc (msec) values with parameters for all sub-groups. Parameters
QTc Formula
Slope
Mean
SD
Min. e Max.
Range
2*SD
95% Confidence Limit
p- value
All samples
Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's Bazett's Fridericia's Van de Water's Hodges' Framingham's
0.110 0.032 0.001 0.229 0.066 0.039 0.280 0.571 0.976 0.330 0.112 0.044 0.020 0.181 0.087 0.104 0.027 0.005 0.224 0.062 0.121 0.041 0.007 0.240 0.075 0.123 0.048 0.018 0.221 0.086 0.093 0.014 0.018 0.224 0.048
272 248 243 286 270 287 250 239 325 276 266 245 242 277 267 272 248 243 286 270 270 245 239 289 268 266 244 240 279 266 275 249 242 295 272
29 24 21 33 22 27 24 19 23 19 23 20 17 23 19 29 25 21 33 22 22 17 14 29 16 20 15 13 25 15 29 25 21 34 22
225e417 218e391 216e372 226e396 230e393 258e407 223e360 216e327 294e396 257e362 219e417 206e391 208e372 220e393 225e393 225e417 218e391 216e372 226e396 230e393 219e358 206e312 208e285 220e373 225e323 222e315 206e274 208e267 226e357 229e294 219e417 215e391 216e372 220e396 225e393
192 173 156 170 163 149 137 111 102 106 198 185 164 173 168 192 173 155 170 163 139 106 77 153 98 93 68 59 131 64 198 176 155 176 168
329 297 284 351 314 341 298 277 371 313 313 286 276 323 305 330 298 285 352 315 314 279 267 346 300 305 274 265 329 296 334 299 285 363 316
0.13930 0.05848 0.02250 0.25230 0.08917 0.21150 0.09009 0.00862 0.52300 0.07950 0.15140 0.07985 0.05208 0.21170 0.11880 0.15390 0.07178 0.03451 0.26290 0.10120 0.15660 0.07431 0.03512 0.26740 0.10180 0.15700 0.07910 0.04587 0.24820 0.11270 0.14220 0.05889 0.01982 0.26290 0.08616
<0.0001 <0.0295 0.9365 <0.0001 <0.0001 0.8634 0.2670 0.0633 <0.0001 0.3253 <0.0001 0.0212 0.2124 <0.0001 <0.0001 <0.0001 0.2560 0.8178 <0.0001 0.0020 <0.0001 0.0125 0.5682 <0.0001 <0.0001 <0.0001 0.0034 0.1774 <0.0001 <0.0001 0.0003 0.5491 0.3417 <0.0001 0.0137
Heart Rate >120 (n ¼ 41)
Heart Rate 120 (n ¼ 142)
Male (n ¼ 91)
Female (n ¼ 92)
Age 15e35 Months (n ¼ 88)
Age >35 Months (n ¼ 95)
to 0.07919 to 0.00330 to 0.02441 to-0.20480 to-0.04231 to 0.25110 to 0.31630 to 0.30430 to 0.20530 to 0.23370 to 0.07288 to 0.00689 to 0.01146 to 0.15020 to 0.05530 to 0.05488 to 0.01941 to 0.04357 to 0.18420 to 0.02333 to 0.08493 to 0.00915 to 0.01942 to 0.21170 to 0.04716 to 0.08900 to 0.01619 to 0.00864 to 0.19530 to 0.05824 to 0.04327 to 0.03155 to 0.05649 to 0.18480 to 0.01002
Key: SD: Standard Deviation; 2SD: 2SD threshold; CL: Confidence limit; Min.: Minimum; Max.: Maximum.
Fig. 1. Uncorrected QT and RR interval.
while, a broad range of QTc values was obtained amongst females and age (15e35 months). The Bazett's formula gave a broad range of QTc values across heart rates (120 and > 120), age (>35 months) and amongst males. In p-value evaluation for all samples analysis, we found that the correction formulae of Bazett, Fridericia, Hodges, and Framingham, the slopes of the QTc verses RR interval relationships were all highly statistically significantly different from zero and therefore were not considered to have adequately corrected for heart rate changes;
while, the correction formula that was considered to have most adequately normalized the QT interval was that of Van de Water's formula, p ¼ 0.9365 because it was not statistically significant different from zero. When the data was divided into subgroups, only one correction formula (Hodges's formula) was found to be highly statistically significant different from zero (p < 0.0001) across the heart rate (>120) and remaining four formulae were not statistically significant different from zero, hence it was considered to have adequately normalized the QT interval, but among them Bazett's formula (p value ¼ 0.8634) was considered to most adequately normalized the QT interval compared to other formulae. In case of sex (male), and age (>35), three formula (Bazett, Hodges, and Framingham formulae) were statistically significant different from zero and remaining two formulae (Van de water and Fridericia formulae) were not statistically significant different from zero, in which Van de Water's formula (p ¼ 0.8178) for male and Fridericia's formulae (p ¼ 0.5491) were considered to most adequately normalized the QT interval compared to other formulae in case of age (>35).
4. Discussion The International Conference on Harmonisation (ICH) recommends that cardiovascular safety pharmacological studies and repeat dose non-rodent toxicity studies be performed, most typically in the dog or primate for any new chemical entity (ICH M3 (R2), 2009). Recording the surface electrocardiogram has been the basic gold standard for determining effects of pharmaceutical compounds on cardiac electrophysiology. A high correlation is
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
121
Fig. 2. Corrected QT intervals for different formulae.
found between action potential prolongation in dog purkinje fibers and increasing QT interval in humans with positive control pharmaceuticals (Gintant et al., 2001). It is important to accurately evaluate the potential of all new candidate drugs to prolong the QT interval preclinical studies (Hammond et al., 2001). Studies have found a negative correlation between HR and QT interval (Spence et al., 1998). Such findings highlight the importance of evaluation of QT interval, by using a correction formula, as variations in HR may mask abnormalities in QT interval. It is difficult to evaluate the clinical and biological significance of minor QT changes, even when they are statistically significant because of the numerous sources of inaccuracies with regard to QT measurement (Moss, 1999). Differences in HR can be a consequence
not only of autonomic conditioning but also of an external factor, like a drug (Malik, 2004). It is very interesting that even during the highly variable HR of SA in the dog; the QT does not change due to QT “memory.” QT memory holds that QT is determined by the average HR measured from the three to six beats preceding the QT being measured (Hamlin et al., 2004). A number of correction formulae are currently used, most of which were derived for analyzing human data, and some for analyzing anaesthetized animals. However little information is available regarding the correction of QT for changes in HR in conscious restrained beagle dogs in toxicology environment (Hanton et al., 2001). For example, in a recent survey, 41% of the 74 laboratories that responded, were using Bazett's formula despite
122
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
Fig. 5. Fridericia's formula for age >35 months. Fig. 3. Bazett's formula for HR > 120.
Fig. 4. Van de water's formula and individual parameters.
the fact that this formula is generally recognized as being unsuitable for correcting dog QT intervals (Hanton et al., 2001; Haverkamp et al., 2000). Our analysis of in-house generated data supports the previous claims of unsuitability of Bazett's and Fridericia's formulae for correction of QT interval in conscious dogs. Moreover, most studies have been done in respect to increase or decrease heart rate means across the heart range against different QTc formula but in current studies, we have analyzed the appropriate QTc formula for dog by considering all pre dose data and also
evaluated after tracing the data w.r.t heart rate, age and sex (Table 2). Previous work comparing various QT correction formulas has produced some conflicting results. One study has concluded that Bazett's and Fridericia's correction formula should be avoided to correct QT interval duration for change in Heart rate in conscious dogs (Hammond et al., 2001; Hanton et al., 2001) while Van de Water's correction factor can be used in toxicology assessment to correct QT interval duration for increases in HR. Our data supports
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
123
Table 3 Basis for recommendations. Recommendations
Basis
Larger sample size
Larger sample size is more representative of a population. It limits the influence of extreme observations or outliers. A large sample size broadens the range of possible data and forms a better picture for analysis. Non-invasive telemetry system Non-invasive telemetry experiments offer long and continuous recordings and simultaneous average acquisitions. Other advantages being: freely moving animals, no surgery, and low stress levels. Institutions must have own QT correction formula To nullify the variations arising due to difference in methodologies regarding conducting the experiments like with its historical data instrument types (Automatic/manual interpretation), source of animals, laboratory conditions and animal husbandry practices. Positive control To assess the reliability of the correction formulae over a wide range of HR
this claim when all samples were considered but when we traced the data for increase in HR above 120 (beats/minute), Bazett's formula was found to be more appropriate than others. This was may be due to insufficient sample size of HR > 120. Hence, we suggest similar data analysis with higher sample size. Given these inconsistencies, some investigators have suggested that separate rate correction formulae may be appropriate for different HR ranges (Rautaharju et al., 2009). Using control data from beagle dogs used in the toxicology assessment, two correction factors described by Bazett (1992) and Van de Water (1989) were identified to adequately correct QT intervals for changes in HR (i.e. slope of QTc versus RR not significantly different from zero). Although both adequately corrected QT intervals, Van de Water's formula, however, showed a statistically superior correction and therefore was identified for use in the evaluation of compound effects. The influence of sex on the evaluation of QT correction was also assessed and showed that sex did not influence the outcome of the evaluations (Fig. 4, Table 2). However, age had a slight effect on the p-value of the formulae, where Fridericia's formula was more suitable than Bazett's formula (Table 2; Fig. 5). We found that Van de Water's formula provides the most consistent correction across a wide range of heart rates. As it can be inferred from Table 2, slope for Van de Water's formula was 0.001 (closest to zero), 2 SD and range of QTc values were 284 and 155 respectively, which are very less compare to other formulae used in this analysis. These values most definitely derive that Van de Water's formula is the most consistent and has lesser variation than other formulae routinely used. Our above conclusion is supported by p-value obtained by Van de Water's formula, which is 0.9635, highly non-significant compared to other formulae. Although, as described before, when HR was considered for >120 (beats/minute), Bazett's formula obtained p-value of 0.8634, making it more suitable to correct QT intervals. Additionally, this analysis was performed using digital electrocardiography with amplification of the initial ECG data to allow for more precise measurements than manual measurement. This instrument gives a direct value of HR, PR, RR, QRS, QT, and hence it is easier to calculate QTc with different formula with accurate results. A limitation inherent to QT measurement occurs when the T and U wave are superimposed or cannot be separated, or when the T and P waves are superimposed, resulting in an inadvertent manual error in measurements of different intervals and thus in QTc values. The analysis performed in this study was performed in accordance with current Recommendations for the Standardization and Interpretation of the ECG (Rautaharju et al., 2009). In conclusion, Van de Water's QTc formula can be used in toxicology assessment to correct QT interval duration across HR and in cases on increased HR; Bazett's formula also has an important role to play. Data is still need to be collected to cover a wide range of HR and further validation is required due to sample size limitation.
Additionally, the development and introduction of newer technologies such as non-invasive telemetry systems that allows continuous ECG recordings in freely moving dogs in a toxicology environment should significantly enhance the quality, reliability, power and ultimately the value of ECG assessment in dog toxicology studies. Organizations can also conduct studies with a positive control to assess the reliability of the correction formulae over a wide range of HR. The authors also recommend institutions to determine appropriate correction formula using its own Historical QT data from its particular strain of dogs and under determined experimental conditions (Table 3). Acknowledgements The authors would like to thank the Study Directors and Study Personnel involved in the designing, conducting and reporting the studies in current evaluation. All the toxicity studies included under this analysis were financed and performed at Zydus Research Centre, Cadila Healthcare Ltd., Ahmedabad, India. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.yrtph.2017.07.026. Transparency document Transparency document related to this article can be found online at http://dx.doi.org/10.1016/j.yrtph.2017.07.026. References Gintant, G. a, Limberis, J.T., McDermott, J.S., Wegner, C.D., Cox, B.F., 2001. The canine Purkinje fiber: an in vitro model system for acquired long QT syndrome and drug-induced arrhythmogenesis. J. Cardiovasc. Pharmacol. 37, 607e618. http:// dx.doi.org/10.1097/00005344-200105000-00012. Gralinski, M.R., 2003. The dog's role in the preclinical assessment of QT interval prolongation. Toxicol. Pathol. (31 Suppl), 11e16. http://dx.doi.org/10.1080/ 01926230390174887. Griffin, G., Stokes, W.S., Pakes, S.P., Gauthier, C., 2004. The ICLAS/CCAC international symposium on regulatory testing and animal welfare. ATLA Altern. Lab. Anim 32, 707e712. Guideline, Committee for Proprietary Medicinal Products (CPMP) & Committee for Veterinary Medicinal Products. Note for Guidance on Inclusion of Antioxidants and Antimicrobial Preservatives in Medicinal Products. 1997. Retreived from: http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_ guideline/2009/09/WC500003408.pdf Guideline, 2009. ICH Harmonised Tripartite. “Guidance on nonclinical safety studies for the conduct of human clinical trials and marketing authorization for pharmaceuticals M3 (R2).”. In: International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Retreived from: http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/ Guidelines/Multidisciplinary/M3_R2/Step4/M3_R2__Guideline.pdf. Guth, B.D., Germeyer, S., Kolb, W., Markert, M., 2004. Developing a strategy for the nonclinical assessment of proarrhythmic risk of pharmaceuticals due to prolonged ventricular repolarization. J. Pharmacol. Toxicol. Methods 49, 159e169. http://dx.doi.org/10.1016/j.vascn.2004.02.006. Hamlin, R.L., Kijtawornrat, A., Keene, B.W., 2004. How many cardiac cycles must be
124
S. Patel et al. / Regulatory Toxicology and Pharmacology 89 (2017) 118e124
measured to permit accurate RR, QT, and QTc estimates in conscious dogs? J. Pharmacol. Toxicol. Methods 50, 103e108. http://dx.doi.org/10.1016/ j.vascn.2004.03.013. Hammond, T.G., Carlsson, L., Davis, A.S., Lynch, W.G., MacKenzie, I., Redfern, W.S., Sullivan, A.T., Camm, A.J., 2001. Methods of collecting and evaluating nonclinical cardiac electrophysiology data in the pharmaceutical industry: results of an international survey. Cardiovasc. Res. 49, 741e750. http://dx.doi.org/ 10.1016/S0008-6363(00)00310-2. Hanton, G., 2007. Preclinical cardiac safety assessment of drugs. Drugs R. D. http:// dx.doi.org/10.2165/00126839-200708040-00002. Hanton, G., Nahas, K., Priou, C., Rabemampianina, Y., Baneux, P., 2001. The Qt interval in the dog Ecg: importance in preclinical toxicology and relationship to heart rate. Toxicol. Methods 11, 21e40. http://dx.doi.org/10.1080/ 105172301300055133. Harada, T., Ishizaki, F., Hamada, M., Horie, N., Nitta, Y., Nitta, K., Katsuoka, H., Nakamura, S., 2009. Circadian rhythm of heart-rate variability and autonomic cardiovascular regulation in Parkinson's disease. Auton. Neurosci. Basic Clin. 46, 333e340. http://dx.doi.org/10.1016/j.autneu.2010.07.016. Haverkamp, W., Breithardt, G., Camm, J. a, Janse, M.J., 2000. The potential for QT prolongation and proarrhythmia by non-antiarrhythmic drugs: clinical and regulatory implications. Eur. Hear. J. 47, 219e233. http://dx.doi.org/10.1053/ euhj.2000.2249. Holzgrefe, H., Ferber, G., Champeroux, P., Gill, M., Honda, M., Greiter-Wilke, A., Baird, T., Meyer, O., Saulnier, M., 2014. Preclinical QT safety assessment: cross-
species comparisons and human translation from an industry consortium. J. Pharmacol. Toxicol. Methods 69, 61e101. http://dx.doi.org/10.1016/ j.vascn.2013.05.004. Malik, M., 2004. Errors and misconceptions in ECG measurement used for the detection of drug induced QT interval prolongation. J. Electrocardiol. 37, 25e33. http://dx.doi.org/10.1016/j.jelectrocard.2004.08.005. Moss, A., 1999. The QT interval and torsade de pointes. Drug Saf. 21, 5e10. Oliveira, M.S., Muzzi, R.A.L., Muzzi, L.A.L., Cherem, M., Mantovani, M.M., 2014. QT interval in healthy dogs: which method of correcting the QT interval in dogs is appropriate for use in small animal clinics? Pesqui. Vet. Bras. 34, 469e472. http://dx.doi.org/10.1590/S0100-736X2014000500014. Rautaharju, P.M., Surawicz, B., Gettes, L.S., 2009. Aha/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram. Part IV: the ST segment, T and U Waves, and the QT interval a scientific statement from the american heart association electrocardiography and arrhythmias Co. J. Am. Coll. Cardiol. 53, 982e991. http://dx.doi.org/10.1016/j.jacc.2008.12.014. Spence, S., Soper, K., Hoe, C.M., Coleman, J., 1998. The heart rate-corrected QT interval of conscious beagle dogs: a formula based on analysis of covariance. Toxicol. Sci. 45, 247e258. http://dx.doi.org/10.1006/toxs.1998.2529. Tattersall, M.L., Dymond, M., Hammond, T., Valentin, J.P., 2006. Correction of QT values to allow for increases in heart rate in conscious Beagle dogs in toxicology assessment. J. Pharmacol. Toxicol. Methods 53, 11e19. http://dx.doi.org/10.1016/ j.vascn.2005.02.005. Tilley, L.P., 1981. Basic canine and feline electrocardiography. Can. Vet. J. 22, 23e25.