Chest Email Content Delivery
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (13)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zamarrón, C.
Right arrow Articles by Romero, P. V.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zamarrón, C.
Right arrow Articles by Romero, P. V.
(Chest. 2003;123:1567-1576.)
© 2003 American College of Chest Physicians

Utility of Oxygen Saturation and Heart Rate Spectral Analysis Obtained From Pulse Oximetric Recordings in the Diagnosis of Sleep Apnea Syndrome*

Carlos Zamarrón, MD; Francisco Gude, MD; Javier Barcala, MD; Jose R. Rodriguez, MD and Pablo V. Romero, MD

* From the Sleep Unit, Division of Respiratory Medicine (Drs. Zamarrón, Barcala, and Rodriguez), and Clinical Epidemiology Unit (Dr. Gude), Hospital Clínico Universitario, Santiago; and Lung Function Test Laboratory (Dr. Romero), Division of Respiratory Medicine, Ciutat Sanitaria y Universitaria de Bellvitge, Barcelona. Spain.

Correspondence to: Carlos Zamarrón, MD, Servicio de Neumología, Hospital Clínico Universitario de Santiago, C/Choupana s/n 15706, Santiago de Compostela, Spain; e-mail: carlos.zamarron.sanz{at}sergas.es


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Objectives: We prospectively evaluate the spectral characteristics of nocturnal arterial oxygen saturation (SaO2) and heart rate variability obtained from pulse oximetric recording as a diagnostic test for obstructive sleep apnea (OSA).

Subjects and measurements: Three hundred referred outpatients with symptoms compatible with the diagnosis of OSA were studied using nocturnal pulse oximetric recording performed simultaneously with polysomnography. Power spectral analysis of SaO2 and heart rate data were analyzed using fast Fourier transformation of a Hamming-windowed signal.

Design and results: Recording test results were classified as abnormal (suspicion of OSA) if the periodogram showed a peak in the period 30 to 70 s in either of the signals. A normal test result was defined as the absence of this peak in the periodogram in both signals. Two independent observers performed a single-blind evaluation. The total area of the periodogram (STOT), the ratio of the area enclosed in the periodogram within the period 30 to 70 s (S30–70), the ratio of the area enclosed in the periodogram within the period 30 to 70 s with respect to the total area of the periodogram (S), and the peak amplitude of the periodogram in the period 30 to 70 s (PA) were measured in both signals. The presence of a peak in the periodogram in either of the signals has a sensitivity of 94%, a specificity of 82%, a positive predictive value of 87%, and a negative predictive value of 92% with respect to the OSA diagnosis. The patients in the OSA group had higher values for STOT, S30–70, S, and PA than the group without OSA.

Conclusions: SaO2 and heart rate spectral analysis obtained by nocturnal pulse oximetry as well as the identification of a peak within 30 to 70 s in either signal could be useful as a diagnostic technique for patient with OSA.

Key Words: heart rate • obstructive sleep apnea • oximetry • polysomnography • spectral analysis


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Obstructive sleep apnea (OSA) is a respiratory disorder characterized by recurrent airflow obstruction caused by total or partial collapse of the upper airway.1 The "gold standard" for a definitive diagnosis of OSA is polysomnography,2 which is an expensive tool and not widely available. Given the high prevalence of OSA,3 4 its potential importance as a contributing factor to cardiovascular morbidity,5 6 7 8 and the availability of an effective treatment for this disease,9 10 numerous efforts have been undertaken to preselect subjects to undergo further clinical investigation.

OSA is frequently accompanied by repetitive oxygen desaturation that can be useful in its detection.11 12 13 14 15 Furthermore, fluctuations in hemodynamic parameters, such as BP and heart rate, during episodes of OSA are well-known phenomena.16 17 18 19 Spectral analysis of arterial oxygen saturation (SaO2) or heart rate variability have been suggested as potential diagnostic tools in this disease20 21 22 23 24 ; however, some patients may not present variations in SaO2 or heart rate signals. Consequently, the study of just one signal may not be sufficient to give a correct diagnosis. Since pulse oximetry equipment can display both signals, the study of the variations in one signal or another may improve the diagnosis of OSA in patients at no additional cost; however, few studies examine the usefulness of SaO2 and heart rate signals together in the diagnosis of this disease.25 The aim of our study was to prospectively evaluate the spectral characteristics of nocturnal SaO2 and heart rate variability obtained from pulse oximetry recording as an OSA diagnostic test.


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Subjects
Three hundred patients (78% men and 22% women) showing clinical symptoms of OSA (age range, 21 to 84 years; mean ± SD body mass index [BMI], 29.5 ± 5.3) were referred to our sleep clinic by general practitioners and included in this study. All patients were suspected of having OSA because of daytime sleepiness, loud snoring, nocturnal choking and awakenings, cease-breathing events, or all four of these as reported by the patient or a bedmate. The Review Board on Human Studies at our institution approved the protocol, and each patient gave his or her informed consent to participate in the study.

Polysomnography
Sleep studies were carried out in our Sleep Unit, usually from midnight to 8 AM. Patients were prospectively evaluated by a single-night pulse oximetry recording (SaO2 and heart rate) obtained by nocturnal pulse oximetry in conjunction with a simultaneous polysomnographic study. This technique consisted of continuous monitoring using a polygraph (Ultrasom Network; Nicolet; Madison, WI) and included EEG, electrooculogram, chin electromyogram, air flow (three-port thermistor), ECG, and measurement of chest wall movement.

The polysomnographic register was analyzed in periods of 30 s and during stages 1, 2, 3, 4, and rapid eye movement (REM) according to the method of Rechtschaffen and Kales.26 Apnea was defined as the absence of airflow for > 10 s, and hypopnea was defined as the reduction of respiratory flow for at least 10 s accompanied by a >= 4% decrease in the saturation of hemoglobin. The average of apnea-hypopnea index (AHI) was calculated in hourly samples of sleep. In this study, an AHI >= 10 was considered as diagnostic of OSA. If the subject had < 3 h of total sleep, the sleep study was repeated.

Power Spectral Analysis of SaO2 and Heart Rate
SaO2 and heart rate recording was performed with a Criticare 504 oximeter (CSI; Wankeska, WI) with a finger probe and sampled at a frequency of 0.2 Hz (one sample every 5 s). In order to avoid aliasing, according to the Niqvist theorem,27 we tested the system to be sure that spectral power at frequencies near and > 0.1 Hz were negligible. Stored time-domain data were played back into a computer for fast Fourier analysis.

Fast Fourier transform (FFT) of the signal inherently assumes that the data we have are a single period of a periodically repeating waveform with an infinite number of samples; however, we have a finite number of samples of a signal that is randomly cut between two points in time. This originates an artifact known as spectral leakage, which is due to the artifactual discontinuities that appear at the beginning and the end of the signal. To avoid spectral leakage, signals were windowed by multiplying them by a Hamming window:

Then the power spectrum of SaO2 and heart rate was analyzed using the FFT of the Hamming-windowed signal.27 Power spectra show the power density or squared magnitude of the amplitude in each of the frequency components of the signal in the bandwidth defined by the interval:


where fmin is lower frequency boundary, fmax is upper frequency boundary, {Delta}f is the frequency resolution, fs is the frequency of sampling (fs = 1/5 = 0.2 Hz), and N is the total number of points sampled. The periodogram is obtained by substituting frequency by period (1/f) in the power spectrum. Analysis was performed by using a Labview (National Instruments Corporation; Austin, TX)-based software. The power contents of the signal were plotted against the period: T(s) = 1/frequency(Hz).21 24

Procedure
The presence of a peak in the periodogram (30 to 70 s) in SaO2 and heart rate recordings was determined in each study. Furthermore, we measured the total area of the periodogram (STOT), the area enclosed in periodogram within the period 30 to 70 s (S30–70), the ratio of the area enclosed in the periodogram within the period 30 to 70 s with respect to the total area of the periodogram (S), and the peak amplitude of the periodogram in the period 30 to 70 s (PA) [Fig 1 ].



View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1.. Top, a: Typical periodogram of a patient with OSA. Power contents of the signal were plotted against the period: T(s) = 1/frequency(Hz). We observed the presence of a peak in the periodogram (30 to 70 s). Bottom, b: Typical periodogram of patients without OSA. Power contents of the signal were plotted against the period: T(s) = 1/frequency(Hz).

 
Two independent observers, one of them with no medical knowledge, classified oximetry recordings blindly. We considered the classification made by the expert observer to be definitive. The trace was classified as abnormal (suspicion of OSA) in the presence of a peak, regardless of the size of the peak in the periodogram (30 to 70 s), and normal in the absence of such a peak in the periodogram (Fig 1) . Fourteen pulse oximetry recordings were rejected because of artifacts or technical problems.

Spirometry was performed using the conventional Collins spirometer (Warren E. Collins; Braintree, MA). Thirty-nine patients had COPD, and 42 patients had cardiovascular disease. Of the 39 patients with COPD, 9 patients (23%) presented with respiratory failure. According to the Global Initiative for Chronic Obstructive Lung Disease consensus,28 19 patients (48%) could be classified as mild, 14 patients (36%) as moderate, and 6 patients (16%) as severe.

With respect to the 42 patients with cardiovascular illness, they had received a diagnosis of hypertension and were receiving antihypertensive medication (with diuretics, angiotensin-converting enzyme inhibitors, or calcium-channel blockers). Of these patients, 12 had ischemic heart disease and 6 were being treated for cardiac failure with digoxin and diuretic medication. No subjects showed episodes of periodic breathing during sleep. All patients were in clinically stable condition during the sleep study.

None of the subjects showed clinical evidence of autonomic dysfunction or neuropathy. Patients were excluded from the study if they had permanent or paroxysmal atrial fibrillation or a permanent pacemaker.

Average of the SaO2 and Heart Rate Spectral Signal
As the differential frequency in the spectrum depends on the total number of samples, which differs among subjects, the spectra were interpolated to obtain the spectral amplitude at equally distributed frequencies between 0 Hz and 0.1 Hz, with a differential frequency in the spectrum of 4 x 10-5 Hz.29 Amplitudes were then averaged to obtain average SaO2 or heart spectra in patients with OSA and without OSA. This procedure was carried out in 120 consecutive subjects: 65 patients with OSA and 55 patients without OSA.

The average of the SaO2 spectral signal is shown in Figure 2 . A peak is present in the 0.010- to 0.033-Hz range (corresponding to the period 30 to 70 s), while it is absent in the spectra of normal subjects. A similar fact is observed when heart rate is considered instead of SaO2 (Fig 3 ). That is, a peak in the spectrum is observed in the same range of frequencies.



View larger version (43K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2.. Top, a: Average SaO2 spectra in patients with OSA. There is a peak in the 0.010- to 0.033-Hz range (corresponding to the period 30 to 70 s). Bottom, b: Average SaO2 spectra in patients without OSA. No peak in the 0.010- to 0.033-Hz range is present. Ampl = amplitude of the signal spectra.

 


View larger version (40K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3.. Top, a: Average heart rate spectra in patients with OSA. There is a peak in the 0.010- to 0.033-Hz range (corresponding to the period 30 to 70 s). Bottom, b: Average heart rate spectra in patients without OSA. No peak in the 0.010- to 0.033-Hz range is observed. bpm = beats per minute; see Figure 2 legend for expansion of other abbreviation.

 
Data Analysis
Data from the pulse oximetry recording (SaO2 and heart rate) was used to determine the sensitivity, specificity, and likelihood ratios of detecting OSA. We considered the utility of both signals individually and together.

Clinical and spectral characteristics of the patients were expressed as means ± SD, or as percentages (95% confidence intervals [CIs]). A Mann-Whitney test was used for comparison between groups. We considered p < 0.05 to be statistically significant. {kappa} statistics were calculated for the measurement of interobserver and intraobserver agreement.


   Results
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Subject Characteristics
The diagnosis of OSA was confirmed in 169 of 300 patients (56%) included in this study. Anthropometric data, AHI, SaO2, and heart rate spectral analysis characteristics are shown in Table 1 . In both SaO2 and heart rate spectral analysis parameters, the group of OSA patients had significantly higher STOT, S30–70, S, and PA values than the group without OSA.


View this table:
[in this window]
[in a new window]

  		Table 1.. Clinical and Spectral Characteristics of the Patients*

 
SaO2 Recording
The SaO2 recording was normal in 124 patients (41.3%) and abnormal in 176 patients (58.7%). The {kappa} index of interobserver concordance was 0.77 (95% CI, 0.70 to 0.84). The concordance between two interpretations in the same observer was 0.90 (95% CI, 0.86 to 0.94). The presence of a peak in the period within 30 to 70 s in the periodogram had a sensitivity of 90% (95% CI, 84 to 94) and a specificity of 82% (95% CI, 74 to 88) for the OSA diagnosis (Table 2 ).


View this table:
[in this window]
[in a new window]

  		Table 2.. Visual Pattern of SaO2, Heart Rate, and the Combined Spectral Analysis in OSA Diagnosis*

 
Heart Rate Recording
The heart rate recording was normal in 143 patients (47.6%) and abnormal in 157 patients (52.4%). The {kappa} index of interobserver concordance was 0.73 (95% CI, 0.66 to 0.81). The concordance between two interpretations in the same observer was 0.91 (95% CI, 0.87 to 0.94) for heart rate. The presence of a peak in the period within 30 to 70 s in the periodogram had a sensitivity of 85% (95% CI, 79 to 90) and a specificity of 90% (95% CI, 84 to 95) for the OSA diagnosis (Table 2) .

Combined SaO2 and Heart Rate Recording
Combined SaO2 and heart rate recording was normal in 149 patients (49.7%) and abnormal in 151 patients (50.3%). The presence of a peak in the period within 30 to 70 s in the periodogram in either signal had a sensitivity of 94% (95% CI, 89 to 97), a specificity of 82% (95% CI, 75 to 88), a positive predictive value of 87% (95% CI, 81 to 92), and a negative predictive value of 92% (95% CI, 85 to 96) for the OSA diagnosis (Table 2) .

In the COPD group (mean age, 63 ± 13 years; mean BMI, 30.4 ± 5.2; 35 men and 4 women) the diagnosis of OSA was made in 26 patients (66%). For patients with COPD and OSA, mean SaO2 was 89 ± 12% and mean heart rate was 78 ± 14 beats/min. For the patients without OSA, mean SaO2 was 90 ± 8% and mean heart rate was 73 ± 13 beats/min. These findings were not statistically significant; however, patients with OSA have a higher baseline heart rate.

The average of the SaO2 and heart rate spectral signals in the group of patients with COPD and OSA are shown in the Figure 4 . A small peak is present in the 0.010- to 0.033-Hz range of the SaO2 spectral signal. Another spectral component is present in the frequency band below 0.01 Hz, while it is absent in the spectra of subjects without OSA. The same is true of the heart rate signal (Fig 4) .



View larger version (65K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4.. Top left, a: Average SaO2 spectra of patients with COPD and OSA. There is a peak in the 0.010- to 0.033-Hz range (corresponding to the period 30 to 70 s) and another spectral component in frequency band below 0.01 Hz. Bottom left, b: Average SaO2 spectra of patients with COPD but without OSA. No peak in the 0.010- to 0.033-Hz range is observed. Top right, c: Average heart rate spectra of patients with COPD and OSA. There is a peak in the 0.010- to 0.033-Hz range and another spectral component in the frequency band below 0.01 Hz. Bottom right, d: Average heart rate spectra of patients with COPD but without OSA. No peak in the 0.010- to 0.033-Hz range is observed. See Figures 2 , 3 for expansion of abbreviations.

 
In the cardiovascular group (mean age, 59 ± 14 years; mean BMI, 31.1 ± 5.4; 22 men and 20 women), the diagnosis of OSA was made in 20 patients (47%). We did not observe any differences in SaO2 and heart rate values between the patients with cardiovascular disease and OSA and those without OSA. For the patients with OSA, mean SaO2 was 94 ± 3% and mean heart rate was 68 ± 11 beats/min. For patients without OSA, mean SaO2 was 94 ± 2% and mean heart rate was 69 ± 10 beats/min.

The characteristics of misclassified subjects after performing the combined methods are shown in Table 3 . Of the false-negative results, 40% were COPD, while most of the remaining patients were older men or obese. Among the false-positive results, 30 to 50% had an AHI of approximately 10.


View this table:
[in this window]
[in a new window]

  		Table 3.. Patient Characteristics When the Diagnosis Was Incorrect (False-negative and False-Positive)*

 

   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
In this study, we show that SaO2 and heart rate spectral characteristics of patients with OSA significantly differ from those of patients without OSA. Furthermore, we quantify these differences.

Periodicities of ventilation have been found in subjects with and without OSA. These ventilatory oscillations originate phase-lagged changes in SaO2 with the same periodicity, and can therefore be detected by spectral analysis of oximetric signal. Basically, this process can be explained as follows: an apnea is defined as a respiratory arrest lasting >= 10 s; including the awakening response after apnea, the most common minimum cycle length of one apneic episode during non-REM lasts approximately 25 s (0.04 Hz). However, the longest apnea time usually observed in OSA during REM lasts approximately 2 min (0.008 Hz). Therefore, OSA-positive patients can be expected to have a peak in the band of the oximetry periodogram between 0.008 Hz and 0.04 Hz (or between 25 s and 120 s).30

Moreover, during apneas and the subsequent hyperventilation periods and arousals, a succession of bradycardia and tachycardia phases is seen in subjects with an intact autonomic nervous system. Frequency domain analysis has shown that the very low frequency component of the heart rate (0.008 to 0.04 Hz) is increased synchronously with the absence of air exchange, which appears repeatedly at a cycle length of 25 to 120 s. The observance of a very low frequency of heart rate variability can be important for OSA detection.30

We focused our attention on the period between 30 to 70 s because of our findings regarding the spectrum average. We have found that a peak is present in the 0.010- to 0.033-Hz range (corresponding to the period 30 to 70 s) of SaO2 or heart rate periodograms in patients with OSA; however, this peak is absent in patients without OSA.

In previous studies21 24 we demonstrated that spectral analysis and peak detection in the period 30 to 70 s of SaO2 or heart rate signal obtained from nocturnal pulse oximetry could be useful as a first step in the diagnosis of OSA. This study, however, goes further by analyzing the usefulness of both SaO2 and heart rate spectral analysis signals together.

Using a computerized method, we demonstrate that the group with OSA has higher values for STOT, S30–70, S, and PA than the group without OSA. The presence of a peak within the period 30 to 70 s in the SaO2 or in heart rate periodogram has a high sensitivity and specificity in the diagnosis of OSA. Indeed, the likelihood ratio for a positive test result was 5.35 and for a negative test result was 0.07, which is well within the range that some authors interpret as generating moderate shifts in pretest to posttest probability.31

According to our results, the visual inspection of the peak occurring in the period 30 to 70 s either in the SaO2 or heart rate periodogram has a higher level of observer agreement (both with one or with multiple observers) than the values reported in the literature on oximetry temporal recordings. In a previous study by Cooper et al,11 which classifies oximetry recordings according to a "pattern recognition" of repetitive dips of SaO2, the multiple observer agreement was only 66%, while we find an agreement of 77% in the SaO2 periodogram and 73% in the heart rate periodogram. Furthermore, the combined spectral analysis that we use has a greater usefulness than other methods utilizing only one signal (either SaO2 or heart rate).

Some studies we reviewed focus on the SaO2 signal. In 240 consecutive patients, Series et al12 found a sensitivity of the visual inspection of the temporal oximetry recording of 98% for OSA, but specificity was only 48%. Levy et al,13 using a mathematical index to detect changes in SaO2, found a sensitivity of 90% and a specificity of 75%. In a more recent study that used a multiple score index on SaO2 recordings, Olson et al14 obtained a sensitivity of 88%, but a low specificity for OSA (40%). However, a prospective study,15 which excluded patients with COPD, showed that analysis of temporal oximetry recordings in suspected OSA has a sensitivity of 80% and specificity of 89%.

With respect to the utility of heart rate signal in the OSA diagnosis, the spectral analysis of heart rate variability using Holter ECG recordings revealed high accuracy for discriminating between normal and periodic breathing with a sensitivity of 90% and specificity of 77%.20 In a later study,22 time-domain analysis of heart rate variability showed a sensitivity of 90%. In another study,23 receiver operating characteristic curves and logistic regression analysis were applied to analyze parameters of heart rate variability in time domain associated with OSA status. The classification and regression methodology showed a sensitivity of 90% and a specificity of 98% by using different thresholds for the same variable. Nevertheless, our combined approach yields both a high sensitivity (94%) and a high specificity (82%).

Of the 10 patients with OSA and a false-negative diagnosis by pulse oximetry, 8 patients with chronic respiratory failure (40% had severe COPD as classified in the Global Initiative for Chronic Obstructive Lung Disease consensus) presented no periodogram peak in the period 30 to 70 s for both heart rate and SaO2 signals, and thus were classified as negative. A possible explanation for this could be that the longer oxygen saturation-resaturation cycles32 for these patients causes a displacement of the periodogram peak (in the period 100 to 200 s), and that these patients have a higher heart rate (mean heart rate of 90 beats/min), which is less variable and with peaks that are less clear. Furthermore, two patients had AHIs very close to the limit, between 10 and 15. Some authors claim that nocturnal SaO2 alone allows for confident recognition of moderate and severe OSA cases, but it is likely to be inadequate for excluding milder cases in clinical practice.12 Fifty percent of patients showing false-negative results are > 65 years old. It is known that heart rate variability decreases with age in healthy people, suggesting an age-dependent decline in autonomic nervous system activity.33 34

Of the 12 patients without OSA and with false-positive diagnoses obtained using both signals together, 50% had an AHI slightly < 10. Therefore, our study included patients with both mild OSA and severe OSA.

Our study has not been designed specifically for upper airway resistance syndrome. Although this syndrome presents no modifications of oxygen saturation, it is associated with arousals related to respiratory effort.35 Because of this, patients with upper airway resistance syndrome will probably show changes in heart rate, which might show up in spectral analysis. Our sample includes three patients with upper airway resistance syndrome that were classified in the non-OSA group and who presented a periodogram peak displacement outside the period 30 to 70 s in the heart rate signal; however, no changes were observed in the SaO2 periodogram.

To our knowledge, this is the first study to compare and combine spectral analysis of SaO2 and heart rate in the diagnosis of OSA. We used simultaneous recordings of overnight pulse oximetry. Although a previous study by Noda et al25 evaluated the relationship between OSA and cardiac disturbances by using the SaO2 and heart rate variability over 24 h in patients with OSA, their study does not provide data of diagnostic value for OSA.

The FFT is an important tool for digital signal processing of the information collected on SaO2 or heart rate. Information is commonly encoded in the sinusoids that form a signal. The shape of the time-domain waveform is not important in these signals; the key information is the frequency and the amplitude of the component sinusoids. The FFT is a mathematical tool used to extract this information. An additional mathematical tool, a Hamming window, which involves the multiplication of the signal by a smooth curve, is used to reduce spectral noise.

For convenience, we developed a user-friendly software using Labview to apply FFT and put this critical information into a usable format. The resulting information is plotted graphically in terms of amplitude and frequency and is known as power spectral analysis.

From a clinical point of view, the inherent complexity of spectral analysis is completely overcome when this user-friendly software is used. Moreover, this software can be incorporated into the digitized oximetry analysis. As a screening tool for the diagnosis of OSA, pulse oximetry is cost-effective and shows substantial accuracy.36

As Bennett and Kinnear37 indicate, for clinicians under pressure from increasing referrals for OSA investigation, oximetry can provide useful information. According to our results, if the test result is negative, it is unlikely that the patient will receive an OSA diagnosis. Nevertheless, other aspects of the clinical history must be taken into consideration. For example, special caution is necessary with patients with COPD and older patients.

Furthermore, oximetry can be performed either in hospital or at home, and is easily processed by computer to give reproducible and objective results. We think that spectral analysis of nocturnal pulse oximetry recording could be a supplementary method to the conventional indexes, and can be incorporated into the same instruments without any additional costs. Each clinician, however, must evaluate oximetric results in the context of his or her own hospital setting.

In conclusion, we have shown that the SaO2 and heart rate spectral characteristics obtained by nocturnal pulse oximetry in patients with OSA are different from those in patients without OSA. Furthermore, we have determined that the presence of a peak in the periodogram of either signal allows us to distinguish patients with OSA from patients without OSA.


   Footnotes
 
Abbreviations: AHI = apnea-hypopnea index; BMI = body mass index; CI = confidence interval; FFT = fast Fourier transform; OSA = obstructive sleep apnea; PA = peak amplitude of the periodogram in the period 30 to 70 s; SaO2 = arterial oxygen saturation; REM = rapid eye movement; S = ratio of the area enclosed in the periodogram within the period 30 to 70 s with respect to the total area of the periodogram; S30–70 = area enclosed in periodogram within the period 30 to 70 s; STOT =total area of the periodogram

This study was supported by Fondo Investigación Sanitaria grant (96/0811) and Secretaria Xeral de Investigacion e Desenvolvemento grant (PGIDT99PXI90201A).

Received for publication October 2, 2001. Accepted for publication December 17, 2002.


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Remmers, JE, DeGroot, WJ, Saurland, EK, et al (1978) Pathogenesis of upper airway occlusion during sleep. J Appl Physiol 44,931-938[Free Full Text]
  2. Practice parameters for the indications for polysomnography and related procedures: an American Sleep Disorders Association report. Sleep 1990;20,406-422
  3. Young, T, Platt, M, Dempsey, J, et al The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328,1230-1235[Abstract/Free Full Text]
  4. Zamarrón, C, Gude, F, Otero, Y, et al Prevalence of sleep disordered breathing and sleep apnea in 50 to 70 year old individuals: a survey. Respiration 1999;66,317-322[CrossRef][ISI][Medline]
  5. Mooe, T, Rabben, T, Wiklund, U, et al Sleep disordered breathing in women: occurrence and association with coronary artery disease. Am J Med 1996;101,251-256[CrossRef][ISI][Medline]
  6. Peppard, PE, Young, T, Palta, M, et al Prospective study of the association between sleep disordered breathing and hypertension. N Engl J Med 2000;342,1378-1384[Abstract/Free Full Text]
  7. Peker, Y, Hedner, J, Kraiczi, H, et al An independent predictor of mortality in coronary artery disease. Am J Respir Crit Care Med 2000;162,81-86[Abstract/Free Full Text]
  8. Shahar, E, Whitney, CW, Redline, S, et al Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med 2001;163,19-25[Abstract/Free Full Text]
  9. American Thoracic Society. Indications and standards for use of nasal continuous positive airway pressure (CPAP) in sleep apnea syndromes: official statement adopted March 1944. Am J Respir Crit Care Med 1994;150,1738-1745[ISI][Medline]
  10. Muñoz, A, Mayoralas, LR, Barbé, F, et al Long-term effects of CPAP on daytime functioning in patients with sleep apnoea syndrome. Eur Respir J 2000;15,676-681[Abstract]
  11. Cooper, BG, Veale, D, Griffiths,, et al Value of nocturnal oxygen saturation as a screening test for sleep apnoea. Thorax 1991;46,586-588[Abstract]
  12. Series, F, Marc, I, Cormier, Y, et al Utility of nocturnal home oximetry for case finding in patients with suspected sleep apnea hypopnea syndrome. Ann Intern Med 1993;119,449-453[Abstract/Free Full Text]
  13. Levy, P, Pépin, JL, Deschaux-Blanc, C, et al Accuracy of oximetry for detection of respiratory disturbances in sleep apnea syndrome. Chest 1996;109,395-399[Abstract/Free Full Text]
  14. Olson, LG, Ambrogetti, A, Gyulay, SG Prediction of sleep disordered breathing by unattended overnight oximetry. J Sleep Res 1999;8,51-55[CrossRef][ISI][Medline]
  15. Chiner, E, Signes, J, Arrieiro, JM, et al Nocturnal oximetry for the diagnosis of the sleep apnoea hypopnoea syndrome: a method to reduce the number of polysomnographies? Thorax 1999;54,968-971[Abstract/Free Full Text]
  16. Leroy, M, Van Surell, C, Pilliere, R, et al Short term variability of blood pressure during sleep in snorers with or without apnea. Hypertension 1996;28,937-943[Abstract/Free Full Text]
  17. Guilleminault, C, Connolly, S, Winkle, R, et al Cyclical variation of the heart rate in sleep apnoea syndrome. Lancet 1984;21,126-131
  18. Dworschak, M, Maurer, JT, Haschemian, T, et al The use of spectral measures of heart rate variability to differentiate between male snorers and patients with sleep apnoea syndrome. Anaesthesia 2001;56,424-428[CrossRef][ISI][Medline]
  19. Wiklund, U, Olofsson, BO, Franklin, K, et al Autonomic cardiovascular regulation in patients with obstructive sleep apnoea: a study based on spectral analysis of heart rate variability. Clin Physiol 2000;20,234-241[CrossRef][ISI][Medline]
  20. Keyl, C, Lemberger, P, Pfeifer, M, et al Heart rate variability in patients with daytime sleepiness suspected of having sleep apnea syndrome: a receiver-operating characteristics analysis. Clin Sci (Colch) 1997;92,335-343[Medline]
  21. Zamarrón, C, Romero, PV, Rodriguez, JR, et al Oximetry spectral analysis in the diagnosis of obstructive sleep apnea. Clin Sci (Colch) 1999;97,467-473[Medline]
  22. Hilton, MF, Bates, RA, Godfrey, KR, et al Evaluation of frequency and time-frequency spectral analysis of heart rate variability as a diagnostic marker of the sleep apnoea syndrome. Med Biol Eng Comput 1999;37,760-769[ISI][Medline]
  23. Roche, F, Gaspoz, JM, Court Fortune, I, et al Screening of obstructive sleep apnea syndrome by heart rate variability analysis. Circulation 1999;100,1411-1415[Abstract/Free Full Text]
  24. Zamarrón, C, Romero, PV, Gude, F, et al Screening of obstructive sleep apnoea: heart rate spectral analysis of nocturnal pulse oximetric recording. Respir Med 2001;95,759-765[CrossRef][ISI][Medline]
  25. Noda, A, Ito, R, Okada, T, et al Twenty four ambulatory oxygen desaturation and electrocardiographic recording in obstructive sleep apnea syndrome. Clin Cardiol 1998;21,506-510[ISI][Medline]
  26. Rechtschaffen, A, Kales, A A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. 1968 BIS/BRI. Los Angeles, CA:
  27. Stearns, SD, Hush, DR Digital signal analysis 2nd ed. 1990,155-162 Prentice Hall International Editions. Englewood Cliffs, NJ:
  28. Global Initiative for Chronic Obstructive Lung Disease. Global strategy for diagnosis, management, and prevention of chronic obstructive pulmonary disease: HLBI/WHO workshop report. 2002 National Institutes of Health, National Heart, Lung, and Blood Institute. Bethesda, MD: Publication No. 2701
  29. Zamarron, C, Romero, P, Otero, Y, et al Spectral analysis of nocturnal variations of arterial oxygen saturation in patients with sleep apnea syndrome (SAS) [abstract]. Am J Respir Crit Care Med 1997;155,A128
  30. Shiomi, T, Guilleminault, Ch, Sasanabe, R, et al Augmented very low frequency component of heart rate variability during obstructive sleep apnea. Sleep 1996;19,370-377[ISI][Medline]
  31. Jaeschke, R, Guyatt, GH, Sackett, DL, for the Evidence-Based Medicine Working Group. Users’ guides to the medical literature: III. How to use an article about a diagnostic test; B. What are the results and will they help me in caring for my patients? JAMA 1994;271,703-707[CrossRef][ISI][Medline]
  32. Calderon-Osuna, E, Carmona Bernal, C, Arenas Gordillo, M, et al A comparative study with chronic obstructive pulmonary disease with and without obstructive sleep apnea syndrome. Arch Bronchoneumol 1999;35,539-543
  33. Shannon, DC, Carley, DW, Benson, H Aging of modulation of heart rate. Am J Physiol 1987;253,H874-H887
  34. Kuo, TB, Lin, T, Yang, CC, et al Effect of aging on gender differences in neural control of heart rate. Am J Physiol 1999;277,H2233-H2239
  35. Guilleminault, C, Stoohs, R, Clerk, A, et al A cause of excessive daytime sleepiness: the upper airway resistance syndrome. Chest 1993;104,781-787[Abstract/Free Full Text]
  36. Netzer, N, Eliasson, AH, Netzer, C, et al Overnight pulse oximetry for sleep-disordered breathing in adults: a review. Chest 2001;120,625-633[Abstract/Free Full Text]
  37. Bennett, JA, Kinnear, WJ Sleep on the cheap: the role of overnight oximetry in the diagnosis of sleep apnoea hypopnoea syndrome. Thorax 1999;54,958-959[Free Full Text]



This article has been cited by other articles:


Home page
 ChestHome page
C.-C. Hua and C.-C. Yu
Smoothed Periodogram of Oxyhemoglobin Saturation by Pulse Oximetry in Sleep Apnea Syndrome: An Automated Analysis
Chest, March 1, 2007; 131(3): 750 - 757.
[Abstract] [Full Text] [PDF]

Home page
 ChestHome page
K. E. Bloch
Getting the Most Out of Nocturnal Pulse Oximetry
Chest, November 1, 2003; 124(5): 1628 - 1630.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (13)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zamarrón, C.
Right arrow Articles by Romero, P. V.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zamarrón, C.
Right arrow Articles by Romero, P. V.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS