How valuable is P-wave dispersion in the determination of carboxyhemoglobin levels?

Abstract

Objective:

To determine whether or not wave/interval dispersions in electrocardiography (ECG) are increased, and to define whether wave and interval dispersions are correlated with carboxyhemoglobin (COHb) levels.

Methods:

ECG, complete blood count, and biochemical parameters were taken from 87 patients with carbon monoxide (CO) poisoning as well as 90 control patients with similar age, gender, and body mass index distribution. COHb levels were recorded in CO-poisoning patients. The COHb levels and the relationships with ECG parameters were studied.

Results:

P_max, P_min, P_d, PR_max, PR_min, PR_d, QT_max, QT_min, QT_d, cQT_max, cQT_min, cQT_d, T_max, T_min, and T_d in ECG were higher in intoxicated patients than the control group (p < 0.05 for all). Pearson’s correlation analyses showed moderately significant positive correlations between COHb level and P_max (r = 0.224; p = 0.037) and P_d (r = 0.222; p = 0.039). The receiver–operator characteristic (ROC) curve showed that a P_d value of 38 ms determined by ECG separates patients with a COHb ≥ 20% with area under the ROC curve of 0.78 (95%CI = 0.71–0.83), a sensitivity of 67.9% (95%CI = 59.4–75.6), a specificity of 95% (95%CI = 83.0–99.2], a positive predictive value of 97.9% (95%CI = 92.5–99.7), and a negative predictive value of 46.3% (95%CI = 35.3–57.7.)

Conclusion:

A significant increase in wave/interval dispersions in the ECG of CO-poisoning patients compared with controls may show that not only a part is affected but both atrium and the ventricles as a whole are affected by hypoxic ischemia. When COHb levels of the patients are unavailable, P dispersion on ECG may show CO poisoning level of the patient.

Keywords

Carboxyhemoglobin CO poisoning P dispersion electrocardiography carbon monoxide

Introduction

Carbon monoxide (CO) is a colorless, odorless gas.¹ CO poisoning is one of the most common types of poisoning, and it is the leading cause of death by poisoning worldwide.² The symptoms can be minimal, even with extremely high levels of carboxyhemoglobin (COHb), as experiments on animals show.³ In patients suffering from CO poisoning, obtaining COHb levels is extremely important by means of medical and legal aspects. For example, COHb levels >20% is an indication for hyperbaric oxygen (O₂) therapy, and it is also an evidence of threatened life judicially.⁴ However, it is a fact that COHb levels can only be measured in large hospitals.

Dispersion is defined as the difference between the maximum and minimum waves/intervals in electrocardiography (ECG). Nowadays, dispersion is studied in various illnesses. Recently, an increase in corrected QT (cQT) and P dispersion has been reported in patients with both acute and chronic CO exposure.^5,6

Our primary aim in this case–control study was to determine whether or not wave/interval dispersions in ECG are increased. Our secondary aim is to define whether wave and interval dispersions in ECG correlated with COHb levels. In this way, we may define objective ECG criteria of the patients with CO poisoning whose COHb levels are unavailable.

Materials and methods

Study population and protocol

The university ethics committee approved the study, and informed consents were collected from the patients. Personal folders of 334 patients with CO poisoning, who were admitted to the Emergency Department (ED) of Atatürk University Faculty of Medicine between the years 2009 and 2012 (39 months), were cross sectionally examined. Of them, 239 patients were excluded due to unavailable ECG and body mass index (BMI) data. ECG, complete blood count (CBC), and biochemical parameters were taken from 87 patients with CO poisoning as well as 90 control patients with similar age, gender, and BMI distribution. Complete blood counts were evaluated with Beckman Coulter LH 780 hematology analyzer. Biochemical parameters were assessed using Cobas 6000 (Roche). COHb levels were recorded in CO-poisoning patients. The COHb levels and the relationships with ECG parameters were studied. A total of 17 patients with unstandardized ECG and/or unavailable dispersion analyze on ECG were later excluded. The control group consisted of healthy patients older than 16 years with minor trauma such as ankle sprain, finger injuries, and so on who were admitted to ED and volunteered to participate in the study.

The initial blood levels of hematocrit, hemoglobin, partial pressure of carbon dioxide, partial pressure of oxygen, sodium, potassium, chloride, calcium, magnesium, lactate, bicarbonate, base excess, and pH were measured for both patients with CO poisoning and control group. Also, COHb levels were measured for intoxicated patients.

For both the controls and patient groups, the following ECG measurements were interpreted:

Heart rate: (HR, beats/min),

P_max−P_min = P dispersion (P_d, ms)

PR_max−PR_min = PR dispersion (PR_d, ms)

QT_max−QT_min = QT dispersion (QT_d, ms)

Corrected QT_max−corrected QT_min = corrected QT dispersion (cQT_d, ms)

T_max−T_min = T dispersion (T_d, ms)

ECG from both groups was measured with the individual measuring blinded to the clinical exposure by the authors KK and YA.

Inclusion criteria

All the patients were older than 16 years with a BMI between 20 and 30 and in sinus rhythm and none were taking medications that cause ECG abnormalities like antiarrhythmic, antihistamines, tricyclic antidepressants, or antipsychotics.

Exclusion criteria

Subjects with thyroid function abnormalities, systemic or pulmonary hypertension, coronary artery disease, valvular heart disease, hypertrophic cardiomyopathy, patients who had pacemakers, significant anemia, cardiac failure, or abnormal ECG were excluded.

CO measurement

CO was assessed by measuring blood COHb level from arterial blood using the Cobas b 221 Blood Gas System (Roche Diagnostics, Inc., Indianapolis, Indiana, USA).

ECG measurements

A 12-lead surface ECG was obtained from all patients while in the supine position at rest with 1 mV/cm amplitude and 25 mm/s rate. The patients were allowed to breathe freely but asked to refrain from speaking or coughing during recording. Durations, P_d, QT_d, and cQT_d were calculated according to the method followed by Sari et al.⁵ P and/or T wave dispersion was calculated by subtracting the minimum P and/or T wave duration from the maximum P and/or T wave duration. The PR interval was defined as the time from the beginning of the P-wave to the beginning of the QRS complex. We found the shortest and longest PR intervals in the leads and measured the PR_d as the difference between maximum and minimum PR intervals according to the method followed by Shah et al.⁷ QT_d was defined as the difference between the longest and shortest QT intervals; rate correction was performed with the Bazett’s formula: corrected QT_d (cQT_d) = QT_d/√(R–R interval) in milliseconds. In all patients, derivations were excluded if the beginning or the ending of the wave could not be clearly identified. Wave durations were measured using a magnifying lens and a 0.5-mm-scale precision ruler. Intra- and interobserver coefficients of variation (the standard deviation of differences between two observations divided by the mean value and expressed as a percentage) were found to be fewer than 5% for all measurements.

Statistical analysis

Continuous variables are expressed as median and categorical data are expressed as percentages. The student’s t test was used for comparing continuous variables and the χ ² test was used to analyze categorical data. The correlation analyses between continuous variables were performed using the Pearson’s correlation test. A two-tailed p < 0.05 was considered to be statistically significant. A receiver–operator characteristic (ROC) curve analysis was used to determine the predictive accuracy of the P_d and P_max for detecting COHb ≥ 20% in patients with CO poisoning. All statistical studies were carried out with the SPSS program (version 20.0, SPSS Inc., Chicago, Illinois, USA).

Results

A total of 87 patients with acute CO poisoning and 90 control patients with similar age, gender, and BMI distribution were enrolled (Table 1). In the intoxication group, the source of CO were natural gas in 6.9% (n = 6), water heater in 31% (n = 27), stove in 58.6% (n = 51), and fire in 3.4% (n = 3). Patients applied to the emergency service on average 1.5 h after poisoning. ECG was taken within 20 min in the ED. Mean time delay for admitting to hospital was 3.1 ± 1.2 h in patients with CO poisoning. There was no difference between the two groups with respect to laboratory analyses (p > 0.05 for each laboratory parameter mentioned above). All the dispersion measurements in ECG were higher in intoxicated patients than the control group (Table 1).

Table 1.

Comparative analysis of clinical and electrocardiograph of CO poisoning and the control group.

Parameter	COHb poisoning (n = 87)	Control group (n = 90)	p Value
Age	34.38 ± 14.83	36.36 ± 11.48	0.098
Male	48%	52%	0.869
BMI (kg/m²)	24.41 ± 3.36	25.07 ± 4.03	0.669
COHb (%)	21.28 ± 12.26	–	–
HR (beats/min)	88.14 ± 15.28	75.62 ± 6.04	<0.0001
P_max (ms)	113.29 ± 17.53	92.48 ± 6.55	<0.0001
P_min (ms)	55.14 ± 13.65	65.11 ± 6.04	<0.0001
P_d (ms)	57.83 ± 14.08	27.37 ± 4.83	<0.0001
PR_max (ms)	184.29 ± 34.20	191.80 ± 9.61	0.019
PR_min (ms)	163.34 ± 34.08	126.62 ± 6.89	<0.0001
PR_d (ms)	20.95 ± 6.0	65.17 ± 9.64	<0.0001
QT_max (ms)	370.96 ± 24.44	372.40 ± 8.98	0.479
QT_min (ms)	321.31 ± 23.52	336.97 ± 8.15	<0.0001
QT_dis (ms)	49.65 ± 6.74	35.42 ± 5.19	<0.0001
cQT_max (ms)	445.70 ± 20.1	417.64 ± 17.40	<0.0001
cQT_min (ms)	385.92 ± 18.82	377.93 ± 16.29	0.018
cQT_d (ms)	59.78 ± 8.70	39.70 ± 5.88	<0.0001
T_max (ms)	150.82 ± 15.33	153.13 ± 7.11	0.29
T_min (ms)	119.08 ± 15.36	124.24 ± 6.03	0.037
T_d (ms)	31.74 ± 5.8	28.88 ± 4.50	0.001

BMI: body mass index; COHb: carboxyhemoglobin; HR: heart rate; P_max: P-wave maximum dispersion; P_min: P-wave minimum dispersion; P_d: P-wave dispersion, PR_max: PR wave maximum dispersion; PR_min: PR wave minimum dispersion; PR_d: PR wave dispersion; QT_max: QT wave maximum dispersion; QT_min: QT wave minimum dispersion; QT_d: QT wave dispersion; cQT_max: corrected QT wave maximum dispersion; cQT_min: corrected QT wave minimum dispersion; cQT_d: corrected QT wave dispersion; T_max: T wave maximum dispersion; T_min: T wave minimum dispersion; T_d: T wave dispersion; CO: carbon monoxide.

For the intoxicated patients, Pearson’s correlation analyses showed weak positive correlations between COHb level and P_max (r = 0.224; p = 0.037) and P_d (r = 0.222; p = 0.039; Figure 1). There was no correlation between COHb level and HR (r = 0.035; p = 0.751), P_min (r = 0.059; p = 0.586), PR_max (r = 0.138; p = 0.204), PR_min (r = 0.15; p = 0.167), PR_d (r = −0.066; p = 0.544), QT_max (r = −0.072; p = 0.509), QT_min (r = −0.049; p = 0.651), QT_d (r = −0.089, p = 0.414), cQT_max (r = −0.015; p = 0.892), cQT_min (r = 0.006; p = 0.959), cQT_d (r = −0.046; p = 0.670), T_max (r = 0.115; p = 0.28), T_min (r = 0.148; p = 0.170), and T_d (r = −0.088; p = 0.415) on ECG. In the poisoning group, P_max was positively correlated with P_d (p < 0.001), although not with age (p = 0.094). P_d was also not associated with age (p = 0.175). The ROC curve showed that a P_d value of 38-ms determined by ECG separates patients with a COHb ≥ 20% with area under the ROC curve (AUC) of 0.78 95% (95%CI = 0.71–0.83), a sensitivity of 67.9% (95%CI = 59.4–75.6), a specificity of 95% (95%CI = 83.0–99.2), a positive predictive value of 97.9% (95%CI = 92.5–99.7), and a negative predictive value of 46.3% (95%CI = 35.3–57.7; Figure 2). The P_d duration value was found to be above 38 ms in 83 patients (95.4%) with CO poisoning and none in the control group (p < 0.0001). On the ROC curve, the AUC of the P_max with the optimum cut off point of 96 ms for predicting a COHb level of ≥20% was 0.629 (95%CI = 0.553–0.700; z statistic = 2.466; significance level P (area = 0.5) = 0.0136). The sensitivity, specificity, positive and negative likelihood ratios (LR+ and LR−, respectively) of this optimum cut off point was 59.12% (CI% = 50.4–67.4%), 75% (CI% = 58.8–87.3%), 2.36 (CI% = 1.9–3.0), and 0.55 (CI% = 0.3–1.0), respectively.

Figure 1.

Linear positive correlation between blood COHb level and P_d. COHb: carboxyhemoglobin; P_d: P-wave dispersion.

Figure 2.

ROC curve analysis the AUC of the P_d for predicting a COHb level of ≥20%. ROC: receiver–operator characteristics; AUC: area under ROC curve; COHb: carboxyhemoglobin; P_d: P-wave dispersion.

Discussion

To the best of our knowledge, this study presents the largest series in terms of comparison of laboratory and ECG findings of CO-poisoning cases to the control group.^5,6,8 This study revealed two significant outcomes: first, we found that all wave durations and dispersions were significantly prolonged in ECG in adult patients with acute CO poisoning compared with the control group. This finding might indicate that the heart is globally (both atriums and ventricles) affected by hypoxia-related CO poisoning; second, among detected dispersions, atria-originated P_max and P-wave dispersions showed a positive linear relationship with the level of COHb. Here, the question of “What level of CO poisoning is demonstrated by P_max and P_d?” arises. According to the literature, P_max and P_d indicate a COHb level of ≥ 20% because COHb < 20% is anticipated as mild poisoning,^9
–11 and this value is considered as the threshold for hyperbaric O₂ therapy.⁴ Furthermore, a COHb level of ≥20% is also the threshold in terms of occurrence of central nervous system effects. The effects of CO on cognitive performance have generally been equivocal at COHb levels of 5–20%.¹² Seizures occur significantly more often in patients with a COHb level of ≥ 20%.¹³ In this regard, we investigated the efficiency of P_max and P_d in terms of demonstration patients with COHb level of ≥ 20%, and P_d was found to show a 95% specificity in achieving this.

As COHb levels may not correlate with symptoms, clinical condition might mislead physicians to determine the severity of the poisoning.³ Many hospital laboratories cannot measure COHb because they do not have CO oximeters. In such instances, blood samples are often sent to outside laboratories or with a transported patient for measurement at the receiving hospital.¹⁴ Therefore, the ECG that is widely used in health care facilities and even in ambulances can be utilized as a noninvasive, inexpensive, easily accessible tool indicating the severity of CO poisoning. In clinical forensic medical practice, the severity of such poisonings must be reported to the judicial authorities. Assessment of P_d/P_max in ECG provides opportunity to determine the severity of the poisoning even in cases without a measured COHb level, especially in retrospective investigations. In the present study, since the sensitivity and specificity of P_d were found to be significantly higher than P_max for determining a COHb level of ≥ 20%, P_d was anticipated to be more useful in terms of showing COHb levels.

P-wave duration in ECG is related to impulse conduction speed and the atrial dimension in normal individuals.¹⁵ Elongation of the P-wave duration reflects left atrial expansion, left atrial hypertension, and conduction anomalies.¹⁶ P-wave dispersion for normal adults has been reported as 27–30 ms in the literature.^17,18 P_d reflects the heterogeneity of electrical conduction in the atria. Distention of the atria due to pressure, volume overload, electrolyte imbalance, or increase in sympathetic activity is a main cause for the increase in P_d. P_max and P_d increase in the following patient groups: atrial fibrillation,^19,20 acute alcohol intake,¹⁶ hyperthyroidism,²¹ coronary artery disease,²² hypertension,²³ atrial septal defect,²⁴ patients with Sendrom X,²⁵ diastolic dysfunction,²⁶ dilated cardiomyopathy,²⁷ congestive heart failure,²⁸ migraine attacks,²⁹ in children with pulmonary arterial hypertension,³⁰ in the elite athletes,^31,32 and in atrial arrhythmias.⁸

The relationship between P_d and CO poisoning has been previously studied. Sari et al. examined 44 indoor barbecue workers and 47 control subjects. They found that there were significant correlations between COHb level and P_d, QT_max, QT_d, and cQT_d.⁵ Cevik et al. found a correlation between reversible increases in QT_d, cQT_max, cQT_d, P_max, and P_d in the ECG and COHb levels in acute CO-poisoning patients.⁶ Recently, Hanci et al. examined 30 CO-poisoning patients and 30 controls.⁸ They found that in patients with acute CO poisoning, the ECG analyses shows that P_d, cQT, and cQT_d durations were significantly prolonged when compared with the control group. They concluded that patients with acute CO poisoning need close follow-up because of arrhythmias. However, this study seems to be the first for revealing a high specificity correlation between P_d and blood COHb levels.

In previously conducted studies, the reason for high P_d in CO poisoning was attributed to increased adrenergic activity, changed conduction and refractory times, increase in sympathetic activity and myocardial damage, and decrease in left ventricle systolic function.^5,6,8 Gunduz et al. found that P-wave dispersion increases in diastolic dysfunction, and this increase was not related to the severity or cause of diastolic dysfunction.²⁶ This association between the COHb levels and changes in electrical conduction in the atrium of intoxicated patients with diminished diastolic dysfunction (abnormal cQT dispersion) and lower ejection fraction^33,34 can be clarified with this pathophysiology: the left atrium is directly exposed to pressure in the left ventricle that increases with decreasing left ventricle compliance. The left atrium pressure increases to maintain adequate filling³⁵ and the increased atrial wall tension leads to chamber dilatation and stretching of the atrial myocardium. This causes a change in the geometry of the atrial fibrils and also in the left atrium wall leading to a disturbance in the interatrial conduction and an inhomogeneous conduction of the sinus impulse.³⁶

Limitations

As there are studies that show that P-wave duration and P_d are influenced by age, sex, and BMI,^30,37 there are also studies that these were not associated with P-wave parameter values.³⁸ P_d values of >40 ms were found to be correlated with atrial fibrillation, with a sensitivity of 74–83% and specificity of 81–85%.¹⁹ However, the present study revealed higher specific relationship of P_d and CO poisoning. We obtained a higher specificity rate for P_d allegedly because our series comprised young-aged and healthy individuals.

The severity of CO poisoning should not only be judged by the serum level of COHb but also by detailed medical history and physical examination.¹¹

P-wave parameters and P_d may vary due to external influences such as seasonal effects³⁹ or internal influences such as anxiety, sleep deprivation, and autonomic function.⁴⁰

The manual calculation of P_d on paper ECG may have influenced our results. Analyzing the ECG using digitally stored ECGs displayed on a high-resolution computer screen, or a high-resolution digitizing board with a specialized software package might be more consistent and acceptable.⁸

Conclusion

A significant increase in wave/interval dispersions in ECG of CO-poisoning patients compared with the controls may show that not only a part is affected but also both atrium and the ventricles as a whole are affected by hypoxic ischemia. All wave durations and dispersions were significantly prolonged in ECG in adult patients with acute CO poisoning compared with the control group, which might indicate that the heart was globally (both atriums and ventricles) affected by hypoxia-related CO poisoning.

Using ECG, P dispersions assessment is valuable to determine the severity of poisoning, especially in situations where COHb level measurement is unavailable. P_d ≥ 38 ms, in cases where COHb level is ≥20%,have 95% specificity if other conditions that increase P_d has been ruled out.

Footnotes

Conflict of interest

The authors declared no conflicts of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Sedda

Rossi

. Death scene evaluation in a case of fatal accidental carbon monoxide toxicity. Forensic Sci Int 2006; 164: 164–167.

Hampson

. Trends in the incidence of carbon monoxide poisoning in the United States. Am J Emerg Med 2005; 23: 838.

Goldbaum

Ramirez

Absalon

. What is the mechanism of carbon monoxide toxicity? Aviat Space Environ Med 1975; 46: 1289–1291.

Iowa Statewide Poison Control Center. http://www.iowapoison.org/index.asp?pageID=148; updated: October 9.

Sari

Zengin

Ozer

Chronic carbon monoxide exposure increases electrocardiographic P-wave and QT dispersion. Inhal Toxicol 2008; 20: 879–884.

Cevik

Tanriverdi

Delice

Reversible increases in QT dispersion and P wave dispersion during carbon monoxide intoxication. Hong Kong J Emerg Med 2010; 17: 441.

Shah

Rhodes

Wieand

P2-104: Non invasive criteria for diagnosis of fasciculoventricular connections in children. Heart Rhythm 2006; 3: S172–S173.

Hanci

Ayoglu

Yurtlu

Effects of acute carbon monoxide poisoning on the P-wave and QT interval dispersions. Anadolu Kardiyol Derg 2011; 11: 48–52.

Goldfrank

Flomenbaum

Lewis

. Carbon monoxide. In: Tomaszewski

(ed) Goldfrank’s toxicologic emergencies. 7th ed. New York, NY: McGraw-Hill, 2002.

10.

Ernst

Zibrak

. Carbon monoxide poisoning. N Engl J Med 1998; 339: 1603–1608.

11.

Grieb

Simons

Piatkowski

Carbon monoxide intoxication versus myocardial infarction: an easy diagnosis?

Burns 2011; 37: e29–e31.

12.

Laties

Merigan

. Behavioral effects of carbon monoxide on animals and man. Annu Rev Pharmacol Toxicol 1979; 19: 357–392.

13.

Keles

Demircan

Kurtoglu

. Carbon monoxide poisoning: how many patients do we miss? Eur J Emerg Med 2008; 15: 154–157.

14.

Hampson

. Stability of carboxyhemoglobin in stored and mailed blood samples. Am J Emerg Med. 2008; 26: 191–195.

15.

Tukek

Akkaya

Demirel

Effect of Valsalva maneuver on surface electrocardiographic P-wave dispersion in paroxysmal atrial fibrillation. Am J Cardiol 2000; 85: 896.

16.

Uyarel

Ozdol

Karabulut

Acute alcohol intake and P-wave dispersion in healthy men. Anatolian J Cardiol 2005; 5: 289.

17.

Aytemir

Ozr

Atalar

P wave dispersion on 12-lead electrocardiography in patients with paroxysmal atrial fibrillation. PACE 2000; 23: 1109–1112.

18.

Weber

Osswald

Huber

Selective versus non-selective antiarrhythmic approach for prevention of atrial fibrillation after coronary surgery: is there a need for pre-operative risk stratification? A prospective placebo-controlled study using low-dose sotalol. Eur Heart J 1998; 19: 794–800.

19.

Dilaveris

Gialafos

. P-wave dispersion: a novel predictor of paroxysmal atrial fibrillation. Ann Noninvasive Electrocardiol 2001; 6: 159–165.

20.

Dogan

Buyukates

Kandemir

Predictors of atrial fibrillation after coronary artery bypass surgery. Coron Artery Dis 2007; 18: 327–331.

21.

Guray

Yilmaz

P wave duration and P wave dispersion in hyperthyroidism: effect of antithyroid treatment. Arch Turk Soc Cardiol 2004; 32: 556–563.

22.

Dilaveris

Andrikopoulos

Metaxas

Effects of ischemia on QT dispersion during spontaneous anginal episodes. J Electrocardiol 1999; 32: 199–206.

23.

Dilaveris

Gialafos

Chrissos

Detection of hypertensive patients at risk for paroxysmal atrial fibrillation during sinus rhythm by computer-assisted P wave analysis. J Hypertension 1999; 17: 1463–1470.

24.

Guray

Y√Ωlmaz

Evaluation of P wave duration and P wave dispersion in adult patients with secundum atrial septal defect during normal sinus rhythm. Int J Cardiol 2003; 91: 75.

25.

Altun

Erdogan

Tatli

Increased P wave dispersion: a new finding in patients with syndrome X. Canad J Cardiol 2002; 18: 1207.

26.

Gunduz

Binak

Arinc

The relationship between P wave dispersion and diastolic dysfunction. Tex Heart Inst J 2005; 32: 163–167.

27.

Senen

Turhan

Erbay

P-wave duration and P-wave dispersion in patients with dilated cardiomyopathy. Eur J Heart Fail 2004; 6: 567–569.

28.

Kim

The relationship between the left atrial volume and the maximum P-wave and P-wave dispersion in patients with congestive heart failure. Yonsei Med J 2007; 48: 810–817.

29.

Duru

Melek

Seyfeli

QTc dispersion and P-wave dispersion during migraine attacks. Cephalalgia 2006; 26: 672–677.

30.

Sap

Karatas

Altin

Dispersion durations of P-wave and QT interval in children with congenital heart disease and pulmonary arterial hypertension. Pediatr Cardiol 2013; 34: 591–596.

31.

Puerta

Aliz

Lopez-Calleja

Increased P wave dispersion in elite athletes. Indian Pacing Electrophysiol J. 2011; 11: 73–80.

32.

Metin

Yildiz

Bayraktar

Assessment of the P wave dispersion and duration in elite women basketball players. Indian Pacing Electrophysiol J 2010; 10: 10–20.

33.

Kalay

Ozdogru

Cetinkaya

Cardiovascular effects of carbon monoxide poisoning. Am J Cardiol 2007; 99: 322–324.

34.

Teksam

Gumus

Bayrakci

Erdogan

Kale

. Acute cardiac effects of carbon monoxide poisoning in children. Eur J Emerg Med 2010; 17: 192–196.

35.

Greenberg

Chatterjee

Parmley

The influence of left ventricular filling pressure on atrial contribution to cardiac output. Am Heart J 1979; 98: 742–751.

36.

Tsang

Gersh

Appleton

Left ventricular diastolic dysfunction as a predictor of the first diagnosed nonvalvular atrial fibrillation in 840 elderly men and women. J Am Coll Cardiol 2002; 40: 1636–1644.

37.

Magnani

Johnson

Sullivan

P-wave indices: derivation of reference values from the Framingham heart study. Ann Noninvasive Electrocardiol 2010; 15: 344–352.

38.

Nussinovitch

. Meta-analysis of P-wave dispersion values in healthy individuals: the influence of clinical characteristics. Ann Noninvasive Electrocardiol 2012; 17: 28–35.

39.

Kose

Aytemir

Can

Seasonal variation of P-wave dispersion in healthy subjects. J Electrocardiol 2002; 35: 307–311.

40.

Sari

Davutoglu

Ozbala

Acute sleep deprivation is associated with increased electrocardiographic P-wave dispersion in healthy young men and women. Pacing Clin Electrophysiol 2008; 31: 438–442.