Blood pressure is one of the most important healthcare indices that can be measured using noninvasive methods, and as for noninvasive blood pressure measurements, the auscultatory method is the gold standard. However, because the observer detects the sounds directly through their ear, ambient noise and the hearing level of the measurer affect this method, calling into question the accuracy and efficacy of this method when the observer suffers from hearing loss.
When measuring blood pressure by the auscultatory method, Korotkoff sounds are used as indices of the systolic blood pressure (SBP) and diastolic blood pressure (DBP). Korotkoff sounds have five distinct phases: phase I is composed of tapping sounds, phase II has soft murmur sounds, phase III has louder slapping sounds, phase IV has sudden muffled sounds, and phase V is the disappearance of sound 1. In the auscultatory method, SBP is defined as the pressure value of the cuff when the Korotkoff sound appears for the first time, corresponding to phase I, and the DBP is defined as the pressure value of the cuff when the Korotkoff sounds disappear, corresponding to phase V. If the individual who measures blood pressure using the auscultatory method has impaired hearing, in case of SBP, he/she can miss the first sound of the Korotkoff sound because it generally increases from the first sound as it continues. Therefore, SBP can be underestimated because the cuff is deflating and the cuff pressure is gradually decreasing. In the case of DBP, the sounds of the Korotkoff sound would become smaller gradually; therefore, an observer with hearing loss might overestimate the DBP. One additional consideration is that the auscultatory method is generally known to underestimate SBP and overestimate DBP in relation to pressures measured intra-arterially 2–4; therefore, it is important to gather as accurate measures as possible.
The audible frequency range for humans is usually between 16 and 20 000 Hz. In the hospital, hearing tests are performed from 125 to 8000 Hz, but generally, the 125 Hz frequency band is skipped; therefore, hearing tests are performed above 250 Hz 5. The hearing test is designed to test the hearing of speech and it is sufficient for over 250 Hz only. However, the Korotkoff sounds have frequency bands from 20 to 300 Hz 6. In particular, phases IV and V Korotkoff sounds, which are specifically used to measure diastolic pressure, are in the 100 Hz frequency range or below 6. Therefore, hearing tests that include frequency bands below 250 Hz should be performed. Hearing loss usually occurs at almost all of the audible frequency bands, but in some cases, it occurs at only the frequency band higher than 250 Hz or only the frequency band below 250 Hz. The latter case, where the frequency band of the Korotkoff sounds exists in, is rare. However, in the case of hearing loss over all the frequency bands, it would have a negative influence on the accuracy of the auscultatory method because it includes the Korotkoff sound frequency band.
Allen et al.6 confirmed in their study that from phase III to phases IV and V of the Korotkoff sounds, which correspond to DBP, there is a significant reduction in maximum amplitude, highest frequency, and high-frequency energy. Also, they found that phase I, which is related to SBP, has a higher maximum amplitude, greater highest frequency, and greater high-frequency energy than phases IV and V. These analyses indicate that the probability of missing the Korotkoff sound for DBP is higher than the probability of missing the SBP sound; thus, we anticipated that the inaccuracy of the auscultatory method would be greater for DBP than SBP measurements.
Therefore, the aim of our study is to confirm the relationship between observer hearing level and the accuracy of the blood pressure measurement using the auscultatory method.
Materials and methods
Simulation of hearing loss
To confirm the accuracy of blood pressure measurement according to the observer’s hearing level, several experts (observers) with different hearing levels should simultaneously assess the blood pressure of the same patient. However, it is impractical to find a sufficient number of experts (observers) with different hearing levels. Therefore, we used a hearing loss simulator, which produces an effect that mimics hearing loss in the observer, to attenuate recorded Korotkoff sound data to represent different degrees of hearing loss. Subsequently, the observers with normal hearing assessed these attenuated Korotkoff sounds to compare the differences between blood pressure measurement accuracy at different hearing levels.
To apply the effect of hearing loss on the Korotkoff sound data, we used the processing sequence of combined simulation developed by Moore and Glasberg 7 and Nejime and Moore 8 to simulate the effect of loudness recruitment and threshold elevation with reduced frequency selectivity. To make a hearing loss simulator produce the same results as the one above, we used MATLAB 7.6 Simulink (Mathworks Inc., Natick, Massachusetts, USA) and added a function to modify the frequency bands with decibel sound pressure level (dB SPL) per unit. To attenuate the frequencies where the K sound exists, a frequency band of 20Hz to 200Hz is added. The output waveform has the same data length and size as the input waveform. Notably, there can be little delay between the output waveform and the input waveform, but this problem did not reach a level where it was a problem.
To evaluate the simulation of hearing loss, we used two data sets of Korotkoff sounds. One was provided by the British Hypertension Society (data set A) and the other was acquired through the digital recording system of the sphygmomanometry (data set B).
Data set A
The British Hypertension Society provides a CD-ROM that includes educational videos and supporting materials with simple instructions for auscultatory blood pressure measurement techniques 9,10. It includes 32 recorded Korotkoff sound video tutorials and answers to each practice video exercise (number 9 and 13 data were excluded). To obtain Korotkoff sound data with hearing loss effects, we extracted the audio files from the video clips in wave file format using Adobe Premiere Pro CS5 (Adobe Systems Inc., San Jose, California, USA) and these audio files were placed in the hearing loss simulator to create attenuated sounds. These attenuated audio files were laid over each original video clip and were used to assess the blood pressure in the same way as the original videos. The attenuation process using the hearing loss simulator was applied to all 32 Korotkoff sound video clips (average SBP 164±37 mmHg and average DBP 101±17 mmHg) for all frequency bands from 20 to 5000 Hz, with each original recording corresponding to an attenuated recording with 5, 10, 15, 20, and 25 dB of simulated hearing loss. Therefore, a total of six levels of hearing loss data, from the original data (no attenuation) to 25 dB attenuated data, were used in the experiment. According to the International Organization for Standardization (ISO, 1964 standard), 25 dB or less of hearing loss is classified as normal hearing. We assumed that individuals who measure blood pressure using the auscultatory method in the medical field have normal hearing level; therefore, the data used in the experiment were applied to hearing loss below 25 dB.
Data set B
To add hypertension and hypotension data to the British Hypertension Society tutorial videos in data set A, Korotkoff sounds data acquired through a recording system developed by Hanyang University were used 11. The system records Korotkoff sounds (sampling rate: 3000 Hz) through a microphone and converts sound into a WAVE file. Simultaneously, the cuff pressure (sampling rate: 150 Hz) is measured through a pressure sensor, which creates a text file. These recorded blood pressure data can be assessed multiple times by an observer through an exclusive player installed on a personal computer. The blood pressure data of visitors at Hanyang University Hospital were collected by this recording system. Institutional Review Board approval was obtained for this study, and each participant provided written informed consent (HYI-10-29). During the recording process, two blood pressure measurement experts assessed blood pressures using a stethoscope and an aneroid by the auscultatory method. Each seated participant placed his or her left arm on the table, and the cuff and the bell were positioned at the level of the heart. If the difference between the blood pressure values measured by the two experts exceeded 4 mmHg, the particular data were excluded. Finally, to simulate hearing loss, we applied the same methods to these Korotkoff sound recordings as described for data set A. Overall, we measured 28 individuals (16 men, 12 women, average age 51±16 years, average SBP 132±29 mmHg, average DBP 82±20 mmHg) for data set B.
Five observers who were trained with the British Hypertension Society educational videos were tasked to assess data set A and data set B. All of the observers were hearing tested at frequency bands from 125 to 8000 Hz and they had no hearing loss at any frequency.
The experiments were conducted in a soundproof room and the volume of the sound player was adjusted to exactly the same level for all observers. Data set A was assessed by playing blood pressure video clips on windows media player in Microsoft Windows 7 (Microsoft, Redmond, Washington, USA). Data set B was assessed using PC software that played the recorded Korotkoff sounds. All observers were allowed to replay their videos or sounds during the experiment 12. All of the data were renumbered by randomized sequences to prevent the observers from memorizing the results. Figure 1 shows the procedures for all experiments.
Statistical analyses were carried out using GraphPad Prism 5 for Windows (GraphPad Software, La Jolla, California, USA).
Five observers assessed SBP and DBP from the original data (no attenuation: 0 dB) and from the corresponding attenuated data (5–25 dB). The mean values were calculated from five blood pressure assessment results of each data by five observers and these mean values were averaged by the same attenuated level. The mean differences and the SDs between the original data (0 dB) and the attenuated data (5–25 dB) are shown in Table 1. As the severity of Korotkoff sound attenuation increased, there were greater differences from the original data in both SBP and DBP measures.
Also, greater attenuation of the sounds tended to result in observer underestimation of SBP and overestimation of DBP, with the magnitude of DBP overestimation being greater than the magnitude of the SBP underestimation.
In the case of SBP, when a 10 dB loss was applied to all 60 original videos, observers were could accurately assess 45 of the videos. When a 25 dB loss was applied, only 20 videos were assessed accurately. However, in the case of DBP with 10 dB hearing loss, only 14 videos were assessed accurately and only four were assessed accurately with a 25 dB loss. Overall, the accuracy of the DBP measurements was worse than the SBP at the same level of hearing loss.
The differences in both SBP and DBP between the original recordings and the hearing loss recordings were statistically significant (Fig. 2a and b). In addition, in the SBP measurements of data set B (Fig. 2c), there were significant differences between the original recordings and the 15–25 dB attenuated recordings only, whereas measurements of DBP (Fig. 2d) showed significant differences between the original and the attenuated sound recordings at all hearing levels.
We also combined the measurements from data sets A and B for analysis, and the results are shown in Fig. 2e and f. In the combined data, there were some cases with maximum mean differences between the original recordings and the attenuated recordings of −13.2 mmHg for SBP and −22.8 mmHg for DBP.
In this study, we explored how observer hearing level influences the results of blood pressure measurements using the auscultatory method. As predicted, all observers showed a similar tendency to underestimate SBP and overestimate DBP as the severity of the hearing loss increased. In addition, the magnitude of DBP overestimation was greater than that of the SBP underestimation at the same level of hearing loss. Specifically, in the case of long-lasting, quiet Korotkoff sounds in phase IV or V, the overestimation of DBP was magnified. However, for SBP measurements of attenuated sounds, the first Korotkoff sound in phase I could be detected with certainty in most cases.
According to our results, DBP was overestimated by an average of 5 mmHg for a 25 dB hearing loss. This may seem to be a small difference, but even this slight difference may result in a normotensive individual being misdiagnosed with hypertension, leading to inappropriate treatment and follow-up. Although the average overestimation was 5 mmHg, several DBP measures were overestimated by 10–20 mmHg. These are critical errors that could lead to an incorrect diagnosis from normotension to hypertension and hypotension to normotension. SDs between observers were a maximum of 4.6 mmHg for the same data and the mean differences of these SDs for each hearing loss level between observers were a maximum of 1.21 mmHg. These results indicate that there were no significant differences between observers; therefore, the results of the blood pressure assessments are reasonably coherent, with minimal interobserver differences.
One reason why accurate blood pressure measurements are important was reported by Franklin et al.13. This group found that the risk of coronary heart disease increases with increments in pulse pressure with fixed SBP 13. Therefore, errors in blood pressure measurement by impaired hearing, either underestimation of SBP or overestimation of DBP, can lead to smaller calculated pulse pressures, leading to underestimated risk predictions of coronary heart disease and an inappropriate avoidance of treatment.
When assessing an observer’s ability to hear Korotkoff sounds according to the standards and tests available, we must consider each test and its potential limitations. For example, the ISO standard considers normal hearing to be less than 25 dB of hearing loss, and it is used to calculate the average pure-tone hearing threshold level at 500, 1000, and 2000 Hz, which are the frequencies that are most critical in speech perception. Recently, the American Academy of Otolaryngology – Head and Neck Surgery added 3000 Hz to these three important speech frequency bands 14. These guides indicate that hearing tests and standards are used with speech perception in mind. Therefore, these standards might be inappropriate when considering the ability to hear low-frequency bands like Korotkoff sounds. Although less than 25 dB of hearing loss is classified as normal hearing and does not manifest in daily life, even this slight amount of hearing loss can lead to incorrect blood pressure assessment when using the auscultatory method.
Our study showed that differences in human hearing could affect the accuracy of blood pressure measurements when using the auscultatory method. Therefore, a precise hearing test for the observers who measure blood pressure using the auscultatory method is necessary. Misreading blood pressures because of hearing loss is a realistic concern, and hearing tests that include frequency bands below 250 Hz should be utilized to avoid this problem. If the hearing level of an observer reaches the lower limit of normal hearing, he or she should recognize the possibility of misreading blood pressures. If the observer’s ability to hear is less than the normal hearing threshold, we recommend using an automatic blood pressure measurement device or refraining from measuring blood pressure using the auscultatory method if possible.
This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2012R1A1A2044174). The authors thank Dr Eoin O’Brien for his thoughtful comments, which contributed to this manuscript.
Conflicts of interest
There are no conflicts of interest.
1. Pickering TG, Hall JE, Appel LJ, Falkner BE, Graves J, Hill MN, et al..Recommendations for blood pressure measurement in humans and experimental animals. Part 1: blood pressure measurement in humans. A statement for professionals from the subcommittee of professional and public education of the American Heart Association Council on high blood pressure research.Hypertension2005;45:142–161.
2. Hunyor SN, Flynn JM, Cochineas C.Comparison of performance of various sphygmomanometers with intra-arterial blood-pressure readings.Br Med J1978;2:159–162.
3. Roberts LN, Smiley JR, Manning GW.A comparison of direct and indirect blood-pressure determinations.Circulation1953;8:232–242.
4. Holland WW, Humerfelt S.Measurement of blood pressure: comparison of intra-arterial and cuff values.Br Med J1964;2:1241–1243.
5. Wang B.Idealism versus reality: what’s your preferred hearing screening procedure?Hear Rev2009;16:16–2440.
6. Allen J, Gehrke T, O’Sullivan JJ, King ST, Murray A.Characterization of the Korotkoff sounds using joint time-frequency analysis.Physiol Meas2004;25:107–117.
7. Moore BCJ, Glasberg BR.Simulation of the effects of loudness recruitment and threshold elevation on the intelligibility of speech in quiet and in a background of speech.J Acoust Soc Am1993;94:2050–2062.
8. Nejime Y, Moore BCJ.Simulation of the effect of threshold elevation and loudness recruitment combined with reduced frequency selectivity on the intelligibility of speech in noise.J Acoust Soc Am1997;102:603–615.
9. .Blood pressure measurement: educational video and supporting materials. DVD2007.Leicester, UK:British Hypertension Society.
10. Beevers G, Lip GYH, O’Brien E.Blood pressure measurement: Part II-conventional sphygmomanometry: technique of auscultatory blood pressure measurement.BMJ2001;322:1043–1047.
11. Lee JS, Park DK, Oh HS, Kim IY, Shen DF, Chee YJ.Digital recording system of shygmomanometry.Blood Press Monit2009;14:77–81.
12. O’Sullivan JJ, Allon J, Murray A.A clinical study of the Korotkoff phases of blood pressure in children.J Hum Hypertens2001;15:197–201.
13. Franklin SS, Khan SA, Wong ND, Larson MG, Levy D.Is pulse pressure useful in predicting risk for coronary heart disease?: The Framingham heart study.Circulation1999;100:354–360.
14. Monsell EM, Balkany TA, Gates GA, Goldenberg RA, Meyerhoff WL, House JW.Committee on Hearing and Equilibrium guidelines for the evaluation of results of treatment of conductive hearing loss.Otolaryngol Head Neck Surg1995;113:186–187.