The Laryngoscope
C 2014 The American Laryngological,
V
Rhinological and Otological Society, Inc.
Factors Influencing Sleep Time With Oxygen Saturation Below 90%
in Sleep-Disordered Breathing
Asli Bostanci, MD; Murat Turhan, MD; Selen Bozkurt, MSc
Objectives/Hypothesis: To determine factors influencing sleep time with oxygen saturation below 90% (ST90) in a population referred to a tertiary sleep center for assessment of possible sleep-disordered breathing (SDB).
Study Design: Retrospective review of demographic and polysomnographic data of 731 consecutive patients with suspected SDB.
Methods: Bivariate correlation analyses were performed, and Spearman rho coefficients were calculated. Variables with
a marginal association with ST90 (P <.05) in the bivariate analysis were included into the multiple regression analysis.
Results: The distributions of SDB were as follows: normal/simple snoring, 18.2%; mild obstructive sleep apnea (OSA),
25.6%; moderate OSA, 17.4%; and severe OSA, 38.9%. The univariate analysis revealed a significant correlation between ST90
and age (r 5 0.204), body mass index (BMI) (r 5 0.440), arousal index (r 5 0.754), apnea-hypopnea index (AHI) (r 5 0.761),
mean oxygen saturation (SaO2) (r 5 20.506), and mean O2 desaturation (r 5 0.858) (P <.001 for all parameters). In multiple
regression analysis, age (P <.001), BMI (P 5.040), male gender (P 5.001), AHI (P <.001), mean SaO2 (P <.001), and mean O2
desaturation (P <.001) were found to be independent parameters influencing ST90. Furthermore, these parameters explained
a significant proportion of variance in ST90 (R2 5 0.794, F (6, 729) 5 464.65, P <.001). Although there was a strong correlation between AHI and ST90 (r 5 0.76, P <.001), a large variation of ST90 values was observed, especially within the severe
OSA group.
Conclusions: The study results provide supporting evidence that patients with similar AHI may have quite different values of ST90, and thereby hypoxia. The stratification of patients with OSA according to AHI combined with ST90 may allow better identification of prognostic information.
Key Words: Obstructive sleep apnea, sleep time with oxygen saturation below 90%, hypoxia.
Level of Evidence: 4
Laryngoscope, 125:1008–1012, 2015
INTRODUCTION
Obstructive sleep apnea (OSA) is a common syndrome characterized by repetitive episodic collapse of
the upper airway and intermittent hypoxia during sleep.
Disturbances in gas exchange lead to oxygen desaturation, hypercapnia, and sleep fragmentation, which contribute to the consequences of OSA including metabolic,
neurocognitive, and cardiovascular effects.1
The upper airway caliber is determined by afferent
sensory input to the brainstem respiratory centers and
efferent motor neural output to the upper airway structures.2 Reestablishing airway patency in OSA is
achieved due to arousal. The mucosal sensory receptor
impairment in the upper airway might cause a delayed
end-apneic arousal and extension of the apnea duration.3
The longer the apnea duration is, the deeper the hypoxia
From the Department of Otolaryngology–Head and Neck Surgery
(A.B., M.T.), and the Department of Biostatistics and Medical Informatics
(S.B.), Akdeniz University School of Medicine, Antalya, Turkey.
Editor’s Note: This Manuscript was accepted for publication
August 29, 2014.
The authors have no funding, financial relationships, or conflicts
of interest to disclose.
Send correspondence to Asli Bostanci, MD, Akdeniz University
Hospital, H Blok K: 1, 07070, Konyaalti, Antalya, Turkey.
E-mail: [email protected]
DOI: 10.1002/lary.24942
Laryngoscope 125: April 2015
1008
becomes. The severity of OSA is stratified by the apneahypopnea index (AHI), which represents only the frequency of apneas and hypopneas per hour of sleep
regardless of duration and morphology.4 AHI does not
completely reflect the pathophysiological characteristics
or severity of hypoxia.5 Moreover, patients with a similar
AHI may have different clinical symptoms and
outcomes.6,7
Total sleep time with oxygen saturation below 90%
(ST90) is an objective parameter that can be easily
obtained from polysomnography (PSG). In recent years,
ST90 has gained increasing attention in OSA research
because of its direct relation to the duration and severity
of hypoxia.8–11 In the current study, we aimed to determine factors influencing ST90 in the population referred
to a tertiary sleep center for assessment of possible
sleep-disordered breathing (SDB).
MATERIALS AND METHODS
Study Design and Patients
From November 2011 to February 2014, a total of 731 consecutive patients with suspected SDB, who underwent complete
polysomnographic evaluation at our accredited sleep disorders
center, were included into this retrospective analysis. Demographic and polysomnographic data as well as information
regarding age, gender, body mass index (BMI), total sleep time,
supine sleep time, percentage of supine sleep position,
Bostanci et al.: Factors Influencing ST90 in OSA
nonsupine (lateral and prone) sleep time, percentage of nonsupine sleep position, sleep efficiency, sleep latency, arousal index,
AHI, AHIsupine, AHI rapid eye movement (AHIREM), AHI non–
rapid eye movement (AHINREM), severity of SDB, longest apnea
duration, mean obstructive apnea duration, total apnea duration, lowest O2 saturation (SaO2), mean SaO2, mean O2 desaturation, oxygen desaturation index (ODI), ST90, and percentage
of cumulative time with oxygen saturation below 90% (CT90)
were all recorded following institutional review board approval.
Patients were excluded if they had central sleep apnea
syndrome, previous treatment for SDB by continuous positive
airway pressure, surgery, and/or oral device, age <18 or >70
years, serious cardiovascular disease, chronic obstructive pulmonary disease, asthma, neuropathy or active neurological disease, medications known to affect peripheral nerves, or
malignancy.
PSG (Compumedics E Series Profusion; Compumedics,
Abbotsford, Victoria, Australia) was scored manually based on
American Academy of Sleep Medicine 2007 criteria.4 Apnea was
defined as cessation of airflow for at least 10 seconds with continued effort (obstructive) or lack of effort (central) to breathe.
Hypopnea was defined as a >50% decrease in a valid measure
of airflow without a requirement for associated oxygen desaturation or arousal, and with a lesser airflow reduction in association with oxygen desaturation of >3%, or an arousal for at least
10 seconds. AHI was defined as the number of apnea and hypopnea occurrences per hour. An AHI of <5 was considered as
normal or simple snoring, 5 to 15 as mild OSA, 15 to 30 as moderate OSA, and >30 as severe OSA.
tion, 6%; longest apnea duration, 41.8 seconds; total
apnea duration, 26.7 minutes; ST90, 7 minutes; and
CT90, 1.8%.
Table II presents univariate and multivariate linear
regression analysis of factors influencing ST90. The univariate analysis revealed a significant correlation
between ST90 and age (r 5 0.204), BMI (r 5 0.440),
arousal index (r 5 0.754), AHI (r 5 0.761), mean SaO2
(r 5 20.506), and mean O2 desaturation (r 5 0.858),
(P <.001 for all parameters). Furthermore, there was a
significant, but weak, positive correlation between
supine sleep position and ST90 (r 5 0.100, P 5.008), and
a significant, but quite weak, negative correlation
between nonsupine sleep position and ST90 (r 5 0.099,
P 5.009). However, in multiple linear regression analysis, age (P <.001), BMI (P 5.040), male gender (P 5.001),
AHI (P <.001), mean SaO2 (P <.001), and mean O2 desaturation (P <.001) remained to be independent parameters influencing ST90 after adjustment for other
confounders. Furthermore, these parameters explained a
significant proportion of variance in ST90 (R2 5 0.794, F
(6, 729) 5 464.65, P <.001). In addition, although there
was a strong correlation between AHI and ST90, a great
variation of ST90 values was observed, especially within
the severe OSA group (Fig. 1).
DISCUSSION
Statistical Analysis
The primary outcome was the determination of factors
influencing the ST90. We expressed data as mean, standard
deviation, median, range, and interquartile range for continuous variables. We reported binary variables as counts and percentages. First, bivariate correlation analyses were performed
to assess relationships between ST90 and other variables.
Because normal distribution assumptions were not met, Spearman rho coefficients were calculated. The Mann-Whitney U test
was utilized to evaluate the statistical significance of differences
of ST90 between the gender groups. After checking for multicollinearity (variance inflation factor < 5) and autocorrelation
(Durbin-Watson test), explanatory variables with a marginal
association with the ST90 (P <.05) in the bivariate analysis
were included in the multiple regression analysis. Stepwise
regression was used to select the model that best explained the
relationship between trans-ST90 and the change in explanatory
variables. All P values were two-sided, with the level of significance set at <.05.
All statistical analyses were performed by a professional
biostatistician using IBM SPSS Statistics 20 software (IBM
Corp., Armonk, NY).
RESULTS
Table I presents the characteristics of the patients.
The median age was 48 years (range, 20–70 years), and
most patients were male (78.4%). The distributions of
severity of SDB were as follows: normal/simple snoring,
18.2%; mild OSA, 25.6%; moderate OSA, 17.4%; and
severe OSA, 38.9%. The median AHI, AHIsupine,
AHIREM, AHINREM, arousal index, and ODI were 19,
36.7, 20.2, 18.6, 22.1, and 14 events/hour, respectively.
The median oxymetric values of patients were as follows:
lowest SaO2, 83%; mean SaO2, 95%; mean O2 desaturaLaryngoscope 125: April 2015
OSA is characterized by episodic obstructions of airflow during sleep, with repetitive oscillations in oxyhemoglobin saturation. Conventionally, it was considered
to be an anatomical pathology, because patients with
OSA mostly have a narrow upper airway due to either
increased soft tissue surrounding the airway or
restricted craniofacial bone structure.12,13 However, OSA
can be encountered in patients with a normal anatomy,
but may not be observed in patients with a narrow
upper airway.14 Currently, there is increasing evidence
supporting that respiratory instability, low lung volume,
low arousal threshold, and impaired neural regulation of
breathing are significant contributors to the pathogenesis of apneas and hypopneas.1 Regardless of underlying
etiology, episodic collapse of the upper airway leads to
chronic intermittent hypoxia (CIH), a hallmark of OSA,
which triggers oxidative stress and chronic inflammation.15,16 Consequently, all these pathophysiological processes give rise to detrimental effects on cardiovascular,
neurocognitive, and metabolic functions.1
The gold standard test for the diagnosis of OSA is
the overnight PSG, with the primary outcome measure
of AHI, which only indicates the number of apneas and
hypopneas per hour.4 However, AHI is incapable of
reflecting the actual duration and severity of hypoxia,
and disease outcome. Patients with similar AHI could
have different cessation of breathing and oxygen desaturation characteristics regarding the duration and depth
of events. These differences mostly affect the symptoms
and consequences of the disorder. Mediano et al.
reported that patients with excessive daytime sleepiness
(EDS) had worse nocturnal oxygenation indices and longer apnea duration than those without EDS, despite the
Bostanci et al.: Factors Influencing ST90 in OSA
1009
TABLE I.
Characteristics of Patients.
Variables
No. (%)
Median [Range], {IQR}
Mean 6 SD
Age, yr
48 [20–70], {16}
47.7 6 10.9
Body mass index, kg/m2
Gender, no. (%)
29.8 [20.2–54.7], {5.9}
30.2 6 4.7
Male
573 (78.4)
Female
Total sleep time, min
158 (21.6)
393 [180–537], {70}
387.9 6 56.2
Sleep efficiency, %
88.2 [54.3–99.9], {12.7}
86.1 6 9.3
Sleep latency, min
Arousal index, events/hr
17.5 [1–175], {20}
22.1 [0.3–115.7], {35.9}
24.3 6 22.2
31.0 6 24.4
AHI, events/hr
19.0 [0–115], {39.0}
28.6 6 25.9
AHIsupine, events/hr
AHIREM, events/hr
36.7 [0–120], {55.3}
20.2 [0–112], {44.3}
39.9 6 30.6
28.2 6 26.0
AHINREM, events/hr
18.6 [0–115], {40.8}
28.6 6 26.9
Longest apnea duration, s
41.8 [0–156.2], {34.8}
44.7 6 24.5
Mean obstructive apnea duration, s
Total apnea duration, min
19.6 [0–55.4], {9.5}
26.7 [0–415.6], {83.3}
20.6 6 7.9
66.8 6 87.3
Lowest SaO2, %
83 [38–96], {12}
80.0 6 10.1
Mean SaO2, %
Mean O2 desaturation, %
95 [88–99], {2}
6 [2–26], {4}
95.0 6 1.5
6.8 6 3.9
ODI, events/hr
14.0 [0–106.0], {27.0}
21.3 6 23.4
ST90, min
CT90, %
7.0 [0–427], {40}
1.8 [0–93], {10}
39.6 6 71.3
9.7 6 17.5
Severity of sleep-disordered breathing, no. (%)
Normal/simple snoring
133 (18.2)
Mild OSA
187 (25.6)
Moderate OSA
Severe OSA
127 (17.4)
284 (38.9)
AHI 5 apnea hypopnea index; CT90 5 percentage of cumulative sleep time with oxygen saturation below 90%; IQR 5 interquartile range; NREM 5 non–
rapid eye movement; ODI 5 oxygen desaturation index; OSA 5 obstructive sleep apnea; REM 5 rapid eye movement; SaO2 5 O2 saturation; SD 5 standard
deviation; ST90 5 total sleep time with oxygen saturation below 90%.
fact that neither the AHI, arousal indices, nor overall
sleep architecture were significantly different between
groups.6 In a study evaluating the validity of a new
index considering factors such as the duration and
degree of hypoxia—the so called “integrated area of
desaturation (IAD)”—Asano et al. reported that patients
who experienced cardiovascular events had significantly
higher IAD regardless of AHI.7 Recently, in a similar
TABLE II.
Univariate and Multivariate Linear Regression Analysis of Factors Influencing Sleep Time With Oxygen Saturation Below 90%.
Univariate Analysis
Variables
Median (IQR)
Age, yr
Body mass index, kg/m2
48 (16)
29.8 (5.9)
Gender, male/female, no. (%)
Arousal index, events/hr
AHI, events/hr
Mean SaO2, %
r
0.204
0.440
9 (49)/3 (17)
22.1 (35.9)
19 (39)
95 (2)
Multiple Linear Regression Analysis
P
Coefficients
95% CI
P
<.001
.040
<.001
<.001
0.551
0.595
0.315 to 0.787
0.028 to 1.162
<.001
11.518
5.321 to 17.715
0.754
0.761
<.001
<.001
0.026
0.574
–
0.438 to 0.709
.745
<.001
.001
20.506
<.001
29.665
211.500 to 27.829
<.001
Mean O2 desaturation, %
Supine sleep position, %
6 (4)
43.7 (41.4)
0.858
0.100
<.001
.008
11.978
0.025
11.068 to 12.889
–
<.001
.163
Nonsupine sleep position, %
56.4 (41.7)
20.099
.009
20.021
–
.230
95% CI 5 95% confidential interval; AHI 5 apnea-hypopnea index; IQR 5 interquartile range; SaO2 5 O2 saturation.
Laryngoscope 125: April 2015
1010
Bostanci et al.: Factors Influencing ST90 in OSA
Fig. 1. Distributions of sleep time with oxygen
saturation below 90% (ST90) values within different apnea-hypopnea index (AHI) severity
categories. OSA 5 obstructive sleep apnea.
study with a median follow-up time of 183 months, Kulkas et al. reported that despite there being no significant
differences in AHI, patients with higher duration of
obstruction and desaturation had increased mortality.5
CIH is usually defined as repeated episodes of
hypoxia interspersed with periodic reoxygenations.17
Currently, there is no universally accepted quantitative
clinical test to measure the intensity and/or severity of
CIH due to the lack of a more precise definition. The
ST90 is an objective parameter that represents the duration of nocturnal hypoxia. Li et al.8 first investigated the
clinical value of ST90 in the evaluation of CIH in
patients with OSA and demonstrated that the correlation coefficients of ST90 with AHI and with the Epworth
Sleepiness Scale were both much higher than those of
the lowest SaO2. Later, Li and Jin noted that ST90 was
strongly correlated with AHI and total apnea duration
(r 5 0.770 and 0.776, respectively).9 Recently, Zhang
et al. demonstrated that after adjustment for BMI and
other cardiovascular risk factors, the ST90 was the
strongest independent predictor of high-sensitivity Creactive protein elevation, which is known to be associated with the severity of OSA.10 They concluded that the
severity of OSA should be stratified by combining AHI
and nocturnal CIH variables, such as ST90 and ODI,
instead of AHI alone. Furthermore, in a study of 119
OSA patients who underwent velopharyngeal surgery,
including uvulopharyngopalatoplasty with transpalatal
advancement pharyngoplasty, the ST90 rather than AHI
was reported to be one of the independent predictors of
surgical success.11
The current study is the largest sample to date that
examines the relationship between various polysomnographic variables and ST90 in a population with suspected SDB. Our results demonstrated that ST90 is
independently influenced by age, BMI, male gender,
AHI, mean SaO2, and mean O2 desaturation, each
Laryngoscope 125: April 2015
explaining a significant proportion of variance. The
study results also provide supporting evidence that
patients with similar AHI may have very different values of ST90, and thereby hypoxia.
Potential limitations of our study include its retrospective nature, with inherent problems of selection
bias, an that it is a single-institution analysis, which
could lead to referral bias. A lack of follow-up data and
lack of analysis of other potential confounders, such as
anthropometric measurements, comorbidities, smoking,
and inflammatory markers, could also be considered
potential limitations.
CONCLUSION
The current study contributes to increased concerns
regarding failure of AHI in reflecting the actual severity
of OSA. Based on our results and available data in the
literature, the stratification of patients with OSA according to the AHI combined with ST90 may allow better
identification of prognostic information and selection of
individualized patient-tailored treatment modalities.
Future randomized controlled trials with longer followup are needed to confirm these findings.
BIBLIOGRAPHY
1. Jordan AS, McSharry DG, Malhotra A. Adult obstructive sleep apnoea.
Lancet 2014;383:736–747.
2. Guilleminault C, Ramar K. Neurologic aspects of sleep apnea: is obstructive sleep apnea a neurologic disorder? Semin Neurol 2009;29:368–371.
3. Kimoff RJ, Sforza E, Champagne V, Ofiara L, Gendron D. Upper airway
sensation in snoring and obstructive sleep apnea. Am J Respir Crit Care
Med 2001;164:250–255.
4. Iber C, Ancoli-Israel S, Chesson A, Quan S. The AASM Manual for the
Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications. Westchester, IL: American Academy of Sleep Medicine; 2007.
5. Kulkas A, Tiihonen P, Julkunen P, Mervaala E, T€oyr€
as J. Novel parameters indicate significant differences in severity of obstructive sleep apnea
with patients having similar apnea-hypopnea index. Med Biol Eng Comput 2013;51:697–708.
Bostanci et al.: Factors Influencing ST90 in OSA
1011
6. Mediano O, Barcel
o A, de la Pe~
na M, Gozal D, Agustı A, Barbe F. Daytime
sleepiness and polysomnographic variables in sleep apnoea patients.
Eur Respir J 2007;30:110–113.
7. Asano K, Takata Y, Usui Y, et al. New index for analysis of polysomnography, “integrated area of desaturation”, is associated with high cardiovascular risk in patients with mild to moderate obstructive sleep apnea.
Respiration 2009;78:278–284.
8. Li JR, Sun JJ, Zhang R, Li CF, Hu Y, Sun YM. Clinical value of TS90% in
evaluation of hypoxemia in patients with obstructive sleep apnea/hypoventilation syndrome [in Chinese]. Zhonghua Yi Xue Za Zhi 2005;85:
3115–3117.
9. Li Q, Jin XJ. Correlations between the duration and frequency of sleep
apnea episode and hypoxemia in patients with obstructive sleep apnea
syndrome [in Chinese]. Zhonghua Er Bi Yan Hou Tou Jing Wai Ke Za
Zhi 2009;44:825–830.
10. Zhang XB, Zen HQ, Lin QC, Chen GP, Chen LD, Chen H. TST, as a polysomnographic variable, is superior to the apnea hypopnea index for
evaluating intermittent hypoxia in severe obstructive sleep apnea. Eur
Arch Otorhinolaryngol 2014;271:2745–2750.
Laryngoscope 125: April 2015
1012
11. Zhang J, Li Y, Cao X, et al. The combination of anatomy and physiology in
predicting the outcomes of velopharyngeal surgery. Laryngoscope 2014;
124:1718–1723.
12. Schwab RJ, Gupta KB, Gefter WB, Metzger LJ, Hoffman EA, Pack AI.
Upper airway and soft tissue anatomy in normal subjects and patients
with sleep-disordered breathing. Significance of the lateral pharyngeal
walls. Am J Respir Crit Care Med 1995;152(5 pt 1):1673–1689.
13. Sutherland K, Lee RW, Cistulli PA. Obesity and craniofacial structure as
risk factors for obstructive sleep apnoea: impact of ethnicity. Respirology
2012;17:213–222.
14. Wellman A, Jordan AS, Malhotra A, et al. Ventilatory control and airway
anatomy in obstructive sleep apnea. Am J Respir Crit Care Med 2004;
170:1225–1232.
15. Lavie L. Obstructive sleep apnoea syndrome—an oxidative stress disorder.
Sleep Med Rev 2003;7:35–51.
16. Prabhakar NR, Kumar GK. Oxidative stress in the systemic and cellular
responses to intermittent hypoxia. Biol Chem 2004;385:217–221.
17. Neubauer J. Invited review: physiological and pathophysiological
responses to intermittent hypoxia. J Appl Physiol 2001;90:1593–1599. .
Bostanci et al.: Factors Influencing ST90 in OSA