Diagnostic Values of P-Wave Dispersion to Detect Diastolic Function in Patient with Hypertension

Background: Hypertension is one of the main causes of cardiovascular disease. Patients with hypertension have increase risk of heart failure compared to populations with normal blood pressure. Clinical evidence shows diastolic dysfunction (DD) can lead to heart failure. Diagnostic of DD with echocardiography is important but access to echocardiography machines is limited compared to electrocardiography (ECG). ECG research correlates P-wave dispersion (PWD) with DD. The aim of this study is to determine the value of PWD to diagnose DD in patients with hypertension. Methods: A cross sectional study was conducted in patients with hypertension at Dr. Sardjito Hospital. Patients received echocardiography, ECG, blood pressure measurement and data recording. The diastolic dysfunction was determined based on 2016 ASE/EACVI criteria. We conducted ROC analysis to determine the cut-off point of P-wave dispersion and the area under the curve (AUC) value, and bivariate analysis on demographic and clinical factors related to PWD. Multivariate analysis was performed to determine the independent factors affecting PWD. Results: 113 patients met the criteria of the study subjects, with 47 men (37.2%), mean age 58.32±11.17 years. Thirteen (11.5%) subjects had DD and 37 subjects (32.7%) with increased PWD. Results showed increased PWD above 71.4 m.s with AUC 76.2%, sensitivity 75%, specificity 72.2%, positive predictive value 33.3%, negative predictive value 96%, and accuracy of 72.5% in diagnosing DD. Conclusion: This is the first study to examine the diagnostic value of PWD to detect diastolic function based on 2016 ASE/EACVI criteria. We found PWD above cut-off point 71.4 m.s has a moderate diagnostic value for detecting DD in patients with hypertension.


INTRODUCTION
Hypertension is one of the leading cause of cardiovascular disease.Study in Malaysia showed that the prevalence of hypertension reached 24% and is a major cause of heart failure with high mortality. 1Hypertension causes thickening of the ventricular wall as compensation for increased pressure on the ventricular wall which ultimately results in concentric hypertrophy of the left ventricle and causes diastolic dysfunction. 2In the left ventricular diastolic dysfunction condition, there will be an increase in end diastolic pressure and increase of the dimension of the left atrium.
The increased left atrial pressure and dimension increase the risk of atrial fibrillation, which in turn will increase diastolic dysfunction severity and worsening of symptoms of heart failure 3 .
Diagnosing diastolic dysfunction with echocardiography is important because of the large number of morbidities that can result from poorly handled diastolic dysfunction, but access to echocardiography machines is limited.

ECG examination
A 12 lead EGG recording was done in patients with supine position, paper speed of 50 mm/sec and amplitude of 20 mm/mV using General Electric (GE) Mac i ECG machine.The ECG recording was then embedded in a sequence on paper and then scanned using a HP Deskjet 2520HC scanner with a resolution of 300 dot per inch (DPI) and stored in Joint Photographic Experts Group (JPG) format and then analysed using the ImageJ software on a high resolution computer screen.The value of P-wave dispersion is obtained by substracting the longest duration of the P-wave and the shortest P-wave recorded in twelve leads in a single ECG recording.The P-wave measurement starts from the beginning of the positive deflection until back to the isoelectric line. 5

Echocardiography examination
Echocardiographic examination was performed using the echocardiography machine

Ethical Consideration
This study was ethicaly approved by the ethical committee of medical research Faculty of Medicine,Universitas Gadjah Mada, Yogyakarta and permission from the Dr. Sardjito General Hospital Yogyakarta, Indonesia.

This research was conducted from March
to April 2017 in Dr. Sardjito Hospital.There were 270 subjects with a history of hypertension.The total subjects met the inclusion criteria were 113 subjects.
The ROC curve analysis showed AUC value 76.2% (p=0.003;95% CI 0.620% -0.920%) at P-wave dispersion cut-off point ≥ 71.4 ms with sensitivity 75% and spesificity 72.3%.Area under the curve (AUC) are in the area 70-80% so that the strength of diagnostic test in this study is moderate. 7When we continued to analyze the P-wave dispersion with cut-off point 71. 4   Bivariate analysis was performed on each confounding factor of P-wave dispersion (Table 2).There were significant differences between groups of subjects with ischemic heart disease and those who received therapy of ARB (p <0.05), while subjects with diastolic blood pressure ≥ 90 mmHg, who had a history of smoking, and had beta-blocker therapy reached a value of p <0.25, so further analysis was required with multivariate analysis.
Based on the results of the multivariate analysis, it can be concluded that the use of ARB is an independent variable that affects the P-wave dispersion in this study with p value <0.05.

DISCUSSION
Our study involved 113 subjects with the majority of subjects being female (62.8%) with an average age of 58.32 ± 11.17 years.screens compared to reading using ECG paper using magnifying glass and calipers. 9he AUC value of this research is 76.2% in the range of 70 -80% so that the strength of the diagnostic test in this study is moderate.
Another approach is clinical, which is compared  2001) studies in patients with stable angina. 13other study found no significant differences between the two groups. 11This result is similar to our study that had significant ischemic heart disease on P-wave dispersion in bivariate tests but not significant after multivariate tests.This result is probably due to our exclusion of patients with decreased fraction ejection where most of the causes are ischemic heart disease.
Gender and age did not show statistically significant differences, this result is different from the study of Barutcu et al.(2009) which stated that P-wave dispersion will increase with age, 14 but according toa study from Dagli et al.
(2008) there were no significant differencesof P-wave dispersion in terms of age and sex in hypertensive patients compared with normal patients. 15In our study systolic and diastolic blood pressure variables also did not differ significantly (p=0.946 and 0.052) respectively.
Our study found that betablocker drugs did not affect P-wave dispersion, which is different from previous studies. 19,22,23This finding can be due to differences in the proportion of patients using beta blockers and the type of beta blockers used.The subjects of this study mostly used bisoprolol where the effect of bisoprolol on P-wave dispersion has not been studied previously.

LIMITATIONS
Several limitations should be considered in interpreting our results.The proportion of subjects with diastolic dysfunction is small when compared to the total number of subjects, so the result may not be able to accurately describe the condition in the general population.
An analytic observational study using cross sectional design was conducted in RSUP Dr. Sardjito Hospital, Yogyakarta, Indonesia from March to April 2017.Echocardiography and ECG data were examined on the same day starting from the issuance of ethical clearance until the research sample was fulfilled.Target populations were patients with hypertension who underwent echocardiography as an outpatient.Inclusion criteria included: male or female aged 18-75 years old with diagnosis of hypertension with normal cardiac systolic function, and normokinetics determined by echocardiography, and willing to complete this research.Exclusion criteria included: patients with congenital heart disease, primary valve disease, permanent pacemakers, P-wave dispersion cannot be calculated due to unidentifiedP-wave, atrial fibrillation, artifacts, and/or indeterminate diastolic function based on ASE/EACVI 2016 recomendation.The required sample size was determined to be 81 subjects.Demographic and clinical data include:

GE Vivid 7 6
Pro (GE Vingmed Ultrasound AS, Horten, Norway) and GE Vivid S6 N (GE Vingmed Ultrasound AS, Horten, Norway) in echocardiography division in Dr. Sardjito General Hospital Yogyakarta.The diastolic dysfunction diagnosis used in this study is based on the American Society of Echocardiography 2016 recommendations.Statistical analysis Statistical analysis was performed with IBM ® SPSS ® version 22 for Windows.The numerical variables are presented in the form of the mean±SD (standard deviation) and categorical variables will be presented as a percentage.Categorical variables testing are compared with 2x2 tables using Chi Squared (x 2 ) or Fisher Exact Tests.The cut-off point of P-wave dispersion was determined using the sensitivity-specificity graph from the receiver operator characteristics (ROC) curve.Numerical data are compared by using unpaired Student t-test or Mann Whitney U test based on the normality of data distribution.Confounding variables were analyzed using bivariate analysis followed by multivariate analysis.If there was a variable with a value of p <0.25 in the bivariate test, it was included in the multivariate analysis.The p value <0.05 was considered statistically significant.The assessment of the consistency of ECG readings and the measurement of diastolic function by echocardiography was performed with Kappa, if the coefficient of Kappa > 0.8 with p <0.05 measurement was considered similar.Sensitivity, specificity, positive and negative predictive values were assessed using 2x2 tables.

Figure1.
Figure1.Receiver Operating Curve (ROC) of diagnostic values of P-wave dispersion to detect diastolic function in patients with hypertension

Figure 2 .
Figure 2. Sensitivity-specificity graph to define optimal cut-off point of P-wave dispersion to detect diastolic function in patients with hypertension to the AUC value of the existing diagnostic methods.The study of Tsai et al. (2013) about the correlation of P-wave dispersion with LAVI and diastolic dysfunction shows an increase in P-wave dispersion above 65 m.s has a 61.7% AUC with a sensitivity of 61.8% and a specificity of 56.95 in detecting diastolic dysfunction. 4Another study by Srivastava et al. (2005) on the role of echocardiographic doppler in the assessment of diastolic dysfunction showed the AUC values for each of the E 'lateral, E' medial, E / e 'velocity parameters were 77%, 80%, and 79%. 10 Gunduz et al. (2005) studied the relationship of P-wave dispersion with diastolic dysfunction and compared with the normal population, with 54% of subjects suffering from hypertension with diastolic dysfunction were divided into 3 classes in this study, and it was found that P-wave dispersion increased significantly in patients with diastolic dysfunction (53 ± 9 m.s vs. 43 ± 9; p <0.01), but the difference of P-wave dispersion between the 3 classes of diastolic dysfunction was not significantly different (p> 0.05). 11Another study compared P-wave dispersion in hypertensive patients with and without a paroxysmal AF history found that P-wave dispersion > 44 m.s can distinguish hypertensive patients with and without paroxysmal atrial fibrillation with 77% sensitivity and 62% specificity. 12Research to diagnose diastolic dysfunction by using P-wave dispersion with all subjects who were hypertension patients has never been done beforeusing the ASE/ EACVI 2016 recommendations.To the author best knowledge, this research is the first to apply the new guidelines to use P-wave dispersion to diagnose diastolic dysfunction.Ischemic heart disease has been known to influence P-wave dispersion based on Dilaveris et al. ( Since research subjects continued to receive antihypertensive therapy, the variety of medication used and their effect on P-wave dispersion could not be completely eliminated as confounding variables.CONCLUSION This research is the first study to investigate the diagnostic value of P-wave dispersion to detect diastolic function using ASE/EACVI 2016 recommendations.Our findings demonstrate that P-wave dispersion above cut-off point 71.4 m.s has moderate diagnostic value to detect diastolic dysfunction.

Table 2 .
Bivariate analysis of clinical characteristic and P-wave dispersion in this study was to use the ratio parameter E/e' where the value of ≥ 15 was classified as diastolic dysfunction, with 60% of subjects having

Table 3 .
Comparison of echocardiographic parameter with P-wave dispersion

Table 4 .
Multivariate analysis of factors which affect P-wave dispersion