doi: 10.1161/hc5001.101063
(Circulation. 2001;104:3087.)
© 2001 American Heart Association, Inc.
Clinical Investigation and Reports |
Minimum Bandwidth Requirements for Recording of Pediatric Electrocardiograms
From the Department of Medical Informatics, Faculty of Medicine and Health Sciences, Erasmus University (P.R.R., J.A.K.), and the Department of Pediatric Cardiology, Sophia Children’s Hospital (M.W.), Rotterdam, the Netherlands.
Correspondence to Peter R. Rijnbeek, MSc, Department of Medical Informatics, Faculty of Medicine and Health Sciences, PO Box 1738, 3000 DR Rotterdam, the Netherlands. E-mail Rijnbeek{at}mi.fgg.eur.nl
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background— Previous studies that determined the frequency content of the pediatric ECG had their limitations: the study population was small or the sampling frequency used by the recording system was low. Therefore, current bandwidth recommendations for recording pediatric ECGs are not well founded. We wanted to establish minimum bandwidth requirements using a large set of pediatric ECGs recorded at a high sampling rate.
Methods and Results— For 2169 children aged 1 day to 16 years, a 12-lead ECG was recorded at a sampling rate of 1200 Hz. The averaged beats of each ECG were passed through digital filters with different cut off points (50 to 300 Hz in 25-Hz steps). We measured the absolute errors in maximum QRS amplitude for each simulated bandwidth and determined the percentage of records with an error >25 µV. We found that in any lead, a bandwidth of 250 Hz yields amplitude errors <25 µV in >95% of the children <1 year. For older children, a gradual decrease in ECG frequency content was demonstrated.
Conclusions— We recommend a minimum bandwidth of 250 Hz to record pediatric ECGs. This bandwidth is considerably higher than the previous recommendation of 150 Hz from the American Heart Association.
Key Words: electrocardiography • pediatrics • bandwidth
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
For accurate recording of ECGs, the bandwidth of the recording system is of major importance. The bandwidth, which is defined as the frequency range between low and high frequency cutoffs (-3 dB), should at least extend to the highest frequency component in the ECG signal. Many studies have been performed to determine the frequency content of the adult ECG,1–8 but only few studies addressed the frequency content of the pediatric ECG.9,10
Riggs et al9 studied the frequency spectrum of the ECG of only 8 children and concluded that the vast majority of information is confined to frequencies <100 Hz. However, determining the frequency content of an ECG by inspecting its frequency spectrum is difficult. The ECG spectra appear to be monotonically decreasing, and it is hard to tell which frequency components belong to the ECG and which to noise. Moreover, this approach does not show the effect of reduced bandwidth on wave amplitudes, which from a clinical point of view is the more important information.
Berson et al10 recorded vectorcardiograms from a group of 600 infants. They filtered the waveforms with different low-pass filters and determined the amplitude differences between filtered and unfiltered waves, concluding that a bandwidth of 100 Hz is required to avoid amplitude errors 
10%. However, the original signals were recorded at a sampling rate of 500 Hz. It has been questioned whether this rate is high enough to obtain accurate measurements in pediatric ECGs.1,4,11
The American Heart Association (AHA) recommends 150 Hz as minimum bandwidth and 500 Hz as minimum sampling rate for recording pediatric ECGs, but these recommendations also state that it is unknown how far the bandwidth of systems has to extend due to the limitations of previous studies.12 In the present study, we wanted to determine the minimum bandwidth requirements for recording pediatric ECGs with the use of a large set of ECGs recorded at a high sampling rate.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The study population consisted of 2169 children, with and without cardiac abnormalities, who had been referred to the pediatric cardiology department of the Sophia Children’s Hospital in Rotterdam, the Netherlands. The Table shows the age distribution of the children. For each child, a 12-lead ECG was recorded at a sampling rate of 1200 Hz using a PC-based acquisition system (Cardio Control). The frequency response of the ECG recorder was flat to 320 Hz (-3 dB point). Following common practice in the Department of Pediatric Cardiology in Rotterdam, V3R was used instead of V3 and V7 instead of V5.
|
All ECGs were processed by the Modular ECG Analysis System (MEANS), which has been extensively evaluated, both by its developers13 and by others.14 For each of the 12 leads, the program computes a representative averaged beat, from which ECG measurements are derived. The averaged beats of the 1200-Hz recordings were passed through digital low-pass filters with different -3 dB points, at 50 to 300 Hz in 25-Hz steps, thus simulating recording systems with reduced bandwidths. The filters were developed with the signal processing toolbox of Matlab. They were designed to have a ripple <0.001 dB in the pass-band and an attenuation <-40 dB within 25 Hz of the cutoff point.
To determine the effect of reduced bandwidth on wave amplitudes, we measured the absolute differences between the maximum QRS deflections in the filtered and unfiltered leads for each of the cutoff points and calculated the 95th percentile of these absolute differences. Because we expected the frequency content of the pediatric ECG to decrease with increasing age, ECGs from children <1 year (n=1045) were taken to determine minimum bandwidth requirements. The total population was used to assess the effect of age on the frequency content of the ECG signals. In addition, we calculated the percentage of recordings in which the absolute differences between the maximum deflections of the filtered and unfiltered leads exceeded 25 µV. The 25-µV threshold was chosen because this was considered an amplitude difference still distinguishable by human interpreters from standard paper ECG recordings.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Figures 1 and 2 show the 95th percentiles of the absolute differences between filtered and unfiltered leads for the maximum positive and maximum negative QRS deflections, respectively. The largest absolute differences in both positive and negative deflections are found in lead V4, whereas the smallest differences are seen in lead aVR for positive deflections and in lead I for negative deflections. The clinically important leads V2 and V6 also show large absolute differences in both positive and negative deflections. Figure 3 presents, for each lead, the percentage of recordings and the differences between the maximum positive QRS deflections in the filtered and unfiltered leads exceeding 25 µV. Figure 4 gives these results for maximum negative deflections. The errors in the positive deflections are higher than in the negative deflections, and amplitude errors increase considerably with decreasing bandwidth. In any lead, a bandwidth of 250 Hz yields amplitude errors <25 µV in >95% of the cases (Figures 1 and 3).
|
|
|
|
Figure 5 shows the relationship between age and the 95th percentile of the difference in maximum positive QRS deflections in lead V1 for different bandwidths, illustrating that ECGs have a higher frequency content at younger ages. For example, for children from 0 to 3 months, the effect of a low-pass filter at 100 Hz triples compared with the oldest children in our study population. In V4, the lead that showed the largest amplitude differences, there is a small initial increase of the difference with increasing age up to 
1 year, with a gradual decrease onward (Figure 6). The other leads show comparable patterns of decreasing frequency content with increasing age.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
To establish minimum bandwidth requirements for the accurate recording of pediatric ECGs, we used a large set of pediatric ECGs recorded at a high sampling frequency of 1200 Hz, thus obviating some of the limitations of previous studies.9,10 To our knowledge, this is the first study that demonstrates the effect of bandwidth limitations for all 12 leads of the ECG and that illustrates the influence of age on frequency content. On the basis of our results, we recommend a minimum bandwidth of 250 Hz for recording pediatric ECGs. This bandwidth requirement of 250 Hz is considerably higher than the 150 Hz previously recommended by the AHA.12 With a bandwidth of 150 Hz, 38% of the cases in our study had an error >25 µV in the maximum positive deflection in lead V4. For leads V2 and V6, these percentages are 25% and 23%, respectively. In vectorcardiographic leads, Berson et al10 found amplitude errors >50 µV in 8% of the R-wave amplitudes and 5% of the S-wave amplitudes when using a 150 Hz filter. In our study, 15% of the positive deflections and 7% of the negative deflections in V4 have amplitude errors >50 µV when a 150 Hz filter is used. These differences between the 2 studies may, in part, be explained by the difference in sampling rate (500 versus 1200 Hz) and by the use of different lead systems. Furthermore, the analyses by Berson et al10 were not done on separate leads but on leads X, Y, and Z combined, which is likely to underestimate the effect of filtering for individual leads. More importantly, however, we found a threshold of 25 µV instead of 50 µV is preferable for measuring the effect of reduced bandwidth on signal amplitudes.
We also studied the effect of age on the frequency content of the ECG signals. As shown in Figures 5 and 6, the frequency content gradually decreases from infancy to adulthood. Strictly speaking, for older children, a lower bandwidth would suffice. Our data for children aged 12 to 16 years indicate that the system bandwidth should be 150 Hz to yield amplitude errors < 25 µV in >95% of the cases in this age group. This is close to the 125 Hz recommendation of the AHA for the adult ECG.12 Nevertheless, we recommend using the minimum bandwidth of 250 Hz for the entire pediatric population, because the age range of the patients will often not be known in advance.
An increase in bandwidth will also affect the sampling rate. Shannon’s theorem prescribes a minimum sampling rate that is twice the highest frequency in a signal, which in our case gives a rate of 500 Hz. However, this theorem is valid only for an infinite sampling period and would require a sophisticated but impractical interpolation technique. Therefore, the AHA recommends a sampling rates of 2 or 3 times the theoretical minimum.12 On the basis of this rule of thumb, a sampling frequency of at least 1000 Hz would seem desirable.
|
Acknowledgments |
|---|
The study was supported by the Netherlands Ministry of Economic Affairs.
Received July 5, 2001; revision received October 15, 2001; accepted October 17, 2001.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
-
Barr RC, Spach MS. Sampling rates required for digital recording of intracellular and extracellular cardiac potentials. Circulation. 1977; 55: 40–48.
[Abstract/Free Full Text]
-
Berson AS, Pipberger HV. Electrocardiographic distortions caused by inadequate high-frequency response of direct-writing electrocardiographs. Am Heart J. 1967; 74: 208–218.[Medline]
[Order article via Infotrieve]
-
Berson AS, Ferguson TA, Batchlor CD, et al. Filtering and sampling for electrocardiographic data processing. Comput Biomed Res. 1977; 10: 605–616.[Medline]
[Order article via Infotrieve]
-
Golden DP, Wolthuis RA, Hoffler GW. A spectral analysis of the normal resting electrocardiogram. IEEE Trans Biomed Eng. 1973; 20: 366–372.[Medline]
[Order article via Infotrieve]
-
Pizzuti GP, Cifaldi S, Nolfe G. Digital sampling rate and ECG analysis. J Biomed Eng. 1985; 7: 247–250.[Medline]
[Order article via Infotrieve]
-
Scher AM, Young AC. Frequency analysis of electrocardiograms. Circ Res. 1960; 8: 344–346.
[Abstract/Free Full Text]
-
Yamamoto H, Miyahara H, Domae A. Is a higher sampling rate desirable in the computer processing of the pediatric electrocardiogram? J Electrocardiol. 1987; 20: 321–328.[Medline]
[Order article via Infotrieve]
-
Seegobin RD, Mohamed SA. Sex differences and frequency content of the Frank lead signal-averaged ECG in a normal population. J Electrocardiol. 1996; 29: 105–109.[Medline]
[Order article via Infotrieve]
-
Riggs T, Isenstein B, Thomas C. Spectral analysis of the normal electrocardiogram in children and adults. J Electrocardiol. 1979; 12: 377–379.[Medline]
[Order article via Infotrieve]
-
Berson AS, Lau FY, Wojick JM, et al. Distortions in infant electrocardiograms caused by inadequate high- frequency response. Am Heart J. 1977; 93: 730–734.[Medline]
[Order article via Infotrieve]
-
Garson A. Clinically significant differences between the "old" analog and the "new" digital electrocardiograms. Am Heart J. 1987; 114: 194–197.[Medline]
[Order article via Infotrieve]
-
Bailey JJ, Berson AS, Garson A, et al. Recommendations for standardization and specifications in automated electrocardiography: bandwidth and digital signal processing. Circulation. 1990; 81: 730–739.
[Free Full Text]
-
Van Bemmel JH, Kors JA, Van Herpen G. Methodology of the modular ECG analysis system MEANS. Methods Inf Med. 1990; 29: 346–353.[Medline]
[Order article via Infotrieve]
-
Willems JL, Arnaud P, Van Bemmel JH, et al. A reference data base for multilead electrocardiographic computer measurement programs. J Am Coll Cardiol. 1987; 10: 1313–1321.[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  P. Kligfield, L. S. Gettes, J. J. Bailey, R. Childers, B. J. Deal, E. W. Hancock, G. van Herpen, J. A. Kors, P. Macfarlane, D. M. Mirvis, et al. Recommendations for the Standardization and Interpretation of the Electrocardiogram: Part I: The Electrocardiogram and Its Technology A Scientific Statement From the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society Endorsed by the International Society for Computerized Electrocardiology J. Am. Coll. Cardiol., March 13, 2007; 49(10): 1109 - 1127. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Kligfield, L. S. Gettes, J. J. Bailey, R. Childers, B. J. Deal, E. W. Hancock, G. van Herpen, J. A. Kors, P. Macfarlane, D. M. Mirvis, et al. Recommendations for the Standardization and Interpretation of the Electrocardiogram: Part I: The Electrocardiogram and Its Technology: A Scientific Statement From the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society Endorsed by the International Society for Computerized Electrocardiology Circulation, March 13, 2007; 115(10): 1306 - 1324. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. F Dickinson The normal ECG in childhood and adolescence Heart, December 1, 2005; 91(12): 1626 - 1630. [Full Text] [PDF]
|
|
|
|
 |
|
 P.J. Schwartz, A. Garson Jr, T. Paul, M. Stramba-Badiale, V.L. Vetter, E. Villain, and C. Wren Guidelines for the interpretation of the neonatal electrocardiogram Eur. Heart J., September 1, 2002; 23(17): 1329 - 1344. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||