Introduction
Obstructive sleep apnea (OSA) is estimated to affect
83.7 million adult patients in the United States (1). OSA is associated with several
diseases, such as hypertension, coronary artery disease and
depression, as well as economic burden (2), including a loss in productivity of US
$86.9 billion annually (2). Studies
have shown that OSA severity may be associated with coronary artery
disease, including coronary calcification or atherosclerosis
(3,4). In one study, patients with OSA and an
apnea-hypopnea index (AHI) of ≥15 events/h demonstrated
significantly higher N-terminal pro B-type natriuretic peptide
(275.8 vs. 131.9 pg/ml) and high-sensitivity cardiac troponin T
levels (0.011 vs. 0.008 ng/ml) than patients with an AHI <15
events/h (3).
The American Academy of Sleep Medicine recommends
either in-laboratory polysomnography or a home sleep apnea test
(HSAT) to diagnose OSA (5).
Although HSATs outperform in-laboratory polysomnography in terms of
cost, patient satisfaction and wait times (6), they have some limitations,
particularly regarding scores for respiratory effort-related
arousal (RERA). RERA is defined as a period encompassing a sequence
of breaths of ≥10 sec, with a flattening of the inspiratory portion
of the nasal pressure followed by arousal, that does not meet
criteria for apnea or hypopnea (7,8).
Unlike apnea or hypopnea events, RERA events begin with a partial
narrowing or functional instability of the upper airway leading to
progressive inspiratory effort to maintain adequate ventilation.
This increasing effort results in elevated negative intrathoracic
pressure during inspiration, a condition known as flow limitation.
Cortical arousal in the brain restores muscle tone of the upper
airway to open the passage (5,8). In
patients with OSA, RERA is associated with prevalent hypertension
with a sensitivity of 75% and a specificity of 83% (9).
High RERA score may result in more severe OSA or a
higher respiratory disturbance index (RDI). Because OSA severity
may be misclassified when relying on HSATs alone, undertreatment
with a continuous positive airway pressure machine (CPAP) may
occur, which can result in early cardiovascular diseases and/or
other consequences (10). For
patients who tend to have an RDI higher than AHI, in-laboratory
polysomnography may be preferred, whereas patients with an RDI
equal to AHI may proceed with HSATs to save diagnostic costs.
However, data remain limited about the phenotypes of patients with
OSA who have RDIs higher than AHIs (11). Therefore, the present study aimed to
evaluate whether any clinical factors predict RDIs higher than AHIs
in patients with OSA.
Materials and methods
Study design and population
From October 2024 to June 2025, a retrospective
analytical study was performed at Khon Kaen University Hospital in
Khon Kaen, Thailand. Inclusion criteria were patients aged ≥18
years diagnosed with OSA based on an AHI of ≥5 times/h by
in-laboratory polysomnography, and both AHI and RDI were scored.
Patients who were pregnant or diagnosed with OSA using HSATs were
excluded. The study protocol was approved by Khon Kaen University
Ethics Committee in Human Research (approval no. HE641504).
The clinical features and polysomnography results of
eligible patients were recorded from medical records. Clinical
features included age, sex, comorbidities and physical signs
[weight, height, body mass index (BMI), neck circumference, blood
pressure and snoring, tired, observed apnea, high blood pressure,
body mass index, age, neck circumference, and gender (STOP-BANG)
score (12); polysomnography
results included AHI, RDI and lowest oxygen saturation. Patients
were categorized into two groups: RDI=AHI and RDI>AHI group.
RERA was scored using the American Academy of Sleep Medicine
updated 2007 manual for scoring sleep and associated events
(13). Polysomnography was
performed by a sleep physician who was not blinded to the
participant clinical features.
Sample size calculation
It was predicted that 15% of patients with OSA would
have an RDI higher than AHI. With a confidence interval of 90% and
a margin of error of 7%, the required sample size was 71.
Statistical analysis
Descriptive statistics were used to compare the
clinical features and polysomnography of both groups. Numerical
variables are presented as median (interquartile range), while
categorical variables were reported as number (percentage). Factors
associated with an RDI higher than AHI were computed using logistic
regression analysis and the P-values of studied variables were
calculated using univariate logistic regression analysis. Factors
with P<0.20 according to univariate logistic regression and
those associated with OSA were included in the multivariate
logistic regression analysis (14,15).
Hosmer-Lemeshow χ2 was calculated for the predictive
model to evaluate goodness of fit. All variables with >50%
missing data were excluded from the predictive model (16). The significant numerical predictor
was calculated for sensitivity and specificity using the area under
the receiver operating characteristic (ROC) curve and an
appropriate cutoff point for the numerical predictor was selected
to provide a sensitivity ≥80%. Statistical analyses were performed
in STATA version 18.5 (College Station). P<0.05 was considered
to indicate a statistically significant difference.
Results
Of the 72 patients with OSA who met the inclusion
criteria, 27 (37.50%) had an RDI higher than AHI. Clinical features
of both groups are shown in Tables
I and II. No significant
difference in clinical features emerged between the groups in terms
of age, sex, comorbidity or physical features. The RDI>AHI group
had more patients with obesity (81.48 vs. 60.00%) and a higher body
weight (85 vs. 70 kg) than the RDI=AHI group.
 | Table IClinical features of patients with
obstructive sleep apnea categorized by relative RDI and AHI. |
Table I
Clinical features of patients with
obstructive sleep apnea categorized by relative RDI and AHI.
| Characteristic | RDI=AHI (n=45) | RDI>AHI
(n=27) | P-value |
|---|
| Median age, years
(interquartile range) | 49 (35-60) | 53 (35-61) | 0.780 |
| Male sex (%) | 14 (31.11) | 12 (44.44) | 0.314 |
| Comorbidity (%) | | | |
|
Hypertension | 19 (42.22) | 7 (25.93) | 0.209 |
|
Diabetes | 11 (24.44) | 4 (14.81) | 0.384 |
|
Obesity | 27 (60.00) | 22 (81.48) | 0.071 |
|
Dyslipidemia | 17 (37.78) | 11 (40.74) | 0.808 |
|
Allergic
rhinitis | 12 (26.67) | 5 (18.52) | 0.569 |
|
Asthma | 4 (8.89) | 0 (0.00) | 0.290 |
|
Hypothyroidism | 1 (2.22) | 1 (3.70) | 0.999 |
|
Chronic
kidney disease | 1 (2.22) | 2 (7.41) | 0.552 |
| Table IIPhysical characteristics features of
patients with obstructive sleep apnea categorized by relative RDI
and AHI. |
Table II
Physical characteristics features of
patients with obstructive sleep apnea categorized by relative RDI
and AHI.
| Characteristic | RDI=AHI (n=45) | RDI>AHI
(n=27) | P-value |
|---|
| Weight, kg | 70.00
(64.00-89.00) | 85.00
(70.00-103.00) | 0.053 |
| Height, cm | 159.00
(155.00-165.00) | 163.00
(157.00-171.00) | 0.136 |
| Body mass index,
kg/m2 | 28.04
(24.68-32.05) | 30.04
(26.04-38.51) | 0.132 |
| Neck circumference,
cm | 38.50
(35.00-40.50) | 40.00
(38.00-42.50) | 0.244 |
| Systolic blood
pressure, mmHg | 133.00
(125.00-138.00) | 131.00
(126.00-140.00) | 0.563 |
| Diastolic blood
pressure, mmHg | 77.00
(72.00-90.00) | 80.00
(73.00-84.00) | 0.834 |
| STOP-BANG score | 4.00 (3.00-5.00) | 4.00 (3.00-5.00) | 0.693 |
RDI>AHI group had a higher AHI (40.7 vs. 27.3
times/h) and RDI (53.6 vs. 27.3 times/h) than the RDI=AHI group
(Table III). The lowest oxygen
saturation was comparable between the groups (84 vs. 86%).
| Table IIIPolysomnography results of patients
with obstructive sleep apnea categorized by relative RDI and
AHI. |
Table III
Polysomnography results of patients
with obstructive sleep apnea categorized by relative RDI and
AHI.
| Characteristic | RDI=AHI (n=45) | RDI>AHI
(n=27) | P-value |
|---|
| Apnea-hypopnea index,
times/h | 27.3 (15.2-44.5) | 40.7 (25.1-61.1) | 0.077 |
| Respiratory
disturbance index, times/h | 27.3 (15.2-44.5) | 53.6 (33.7-67.8) | 0.006 |
| Lowest oxygen
saturation, % | 86 (80-89) | 84 (81-88) | 0.828 |
There were five factors in the predictive model for
RDI higher than AHI: Age, sex, hypertension, diabetes, BMI and
STOP-BANG score (Table IV). Only
BMI was independently associated with RDI>AHI, with an adjusted
odds ratio of 1.10 (95% confidence interval: 1.01, 1.19). The
Hosmer-Lemeshow χ2 value of the model was 4.03, which
indicates goodness of fit. The ROC curve of BMI for RDI>AHI was
60.66% (95% confidence interval: 46.53%, 74.78%; Fig. 1). BMI cutoff points for RDI>AHI
are shown in Table V. A BMI of
≥26.03 kg/m2 had a sensitivity of 81.48% and a
specificity of 33.33%.
| Table IVPredictors of presence of respiratory
disturbance index higher than apnea-hypopnea index in patients with
obstructive sleep apnea. |
Table IV
Predictors of presence of respiratory
disturbance index higher than apnea-hypopnea index in patients with
obstructive sleep apnea.
| Predictor | Unadjusted odds ratio
(95% confidence interval) | Adjusted odds ratio
(95% confidence interval) |
|---|
| Age | 1.00 (0.72, 1.04);
0.785 | 1.03 (0.99, 1.07);
0.125 |
| Male sex | 0.56 (0.21, 1.51);
0.256 | 0.45 (0.15, 1.32);
0.147 |
| Hypertension | 0.48 (0.17, 1.36);
0.167 | 0.45 (0.14, 1.42);
0.173 |
| Diabetes | 0.54 (0.15, 1.90);
0.335 | 0.42 (0.09, 1.88);
0.260 |
| Body mass index | 1.05 (0.99, 1.12);
0.122 | 1.10 (1.01, 1.19);
0.022 |
| STOP-BANG score | 1.03 (0.74, 1.44);
0.848 | 0.99 (0.69, 1.43);
0.979 |
| Table VBMI cut-off points of presence of
respiratory disturbance index higher than apnea-hypopnea index in
patients with obstructive sleep apnea. |
Table V
BMI cut-off points of presence of
respiratory disturbance index higher than apnea-hypopnea index in
patients with obstructive sleep apnea.
| BMI,
kg/m2 | Sensitivity, % | Specificity, % |
|---|
| 23.44-26.02 | 85.19 | 11.11 |
| 26.03-30.03 | 81.48 | 33.33 |
| 30.04-32.98 | 51.85 | 62.22 |
| >32.99 | 40.74 | 80.00 |
Discussion
Of all participants, 37.5% had an RDI higher than
AHI; higher BMI is suggestive for RDI >AHI in patients with OSA.
A previous study on the general population showed that the
prevalence of patients with a RERA index ≥5 events/h is 3.8%, with
a RERA index of 7.1 events/h, an RDI of 23.2 events/h and an AHI of
16.1 events/h (11). In the present
study, 37.5% of patients with OSA had a higher RDI than AHI, with
an RDI of 53.6 and AHI of 40.7 events/h. These differences may be
explained by different study populations; the aforementioned study
was a population-based study, whereas the present study was
conducted with patients with OSA. As expected, the severity of OSA
in patients with an RDI higher than AHI was more severe; the median
RDI was 53.6 events/h, categorized as ‘severe OSA’, whereas the
median RDI for individuals with an RDI equal to AHI was 27.3
events/h, classified as ‘moderate OSA’ (5). The RDI>AHI group may have
demonstrated more severe OSA compared with the RDI=AHI group, with
OSA severity potentially ranging from mild to severe and a median
inter-group difference of 26.3 events/h. These findings are
clinical significance, as the severity progression from mild to
severe OSA has important treatment implications-specifically,
severe OSA typically necessitates CPAP therapy (5).
The present study also showed that individuals with
a high BMI, particularly >26.03 kg/m2, may be at risk
of having an RDI higher than AHI, with a sensitivity of 84.84%.
Those patients may benefit from in-laboratory polysomnography more
than HSATs because they may have more severe OSA defined by
in-laboratory polysomnography. These results may differ from a
previous study conducted with obese vs. non-obese patients with OSA
in India (17). In the
aforementioned study, there was a significantly lower proportion of
patients with a RERA index >5 events/h in obese patients than in
the non-obese group (29.16 vs. 52.12%). The aforementioned study
classified obesity as a BMI of 27.5 kg/m2, whereas the
present study adopted a cutoff point of 26.03 kg/m2 for
having RDI higher than AHI. Moreover, the aforementioned study
examined RERA index >5 events/h, whereas the present study
focused on the median RDI index. Further studies are required to
confirm the present results.
BMI and RERA are related. A previous study showed
that patients with OSA with a low arousal threshold have
significantly higher body fat than non-OSA patients based on
visceral fat score (10.95 vs. 6.35) and body fat percentage (28.61
vs. 26.61) (18). The BMI among
patients with a low arousal threshold and OSA was significantly
higher than in the non-OSA group (25.99 vs. 22.74
kg/m2). Those findings may indicate that having high BMI
and OSA may be associated with a low arousal threshold that leads
to higher RERA and RDI.
Although BMI may be a notable predictor of having an
RDI higher than AHI, it has low specificity (33.33%). Therefore,
clinical decision-making should integrate other factors such as
oropharyngeal anatomy. BMI ≥26.03 kg/m2 had a
sensitivity of 81.48%; this cutoff point may be useful in
screening. In patients with suspected OSA and with a BMI <26.03
kg/m2, having RDI higher than AHI is unlikely, and using
HSATs may be justified. By contrast, in-laboratory polysomnography
may be suitable for patients with a BMI ≥26.03
kg/m2.
The present study had limitations. First, the median
BMI of the study population was ~30 kg/m2. Thus, the
present results may not be generalized to other populations such as
non-obese or severely morbid obese patients. Differences in ethnic,
anthropometric, healthcare systems and lack of external validation
may also limit their applicability. Furthermore, variations in
socioeconomic status, the presence of comorbidities, genetic
polymorphisms and phenotypes may prevent generalized
implementation. Second, RERA scoring may be scorer-dependent and
lab-specific, with variability that may affect RERA burden and
validity. An RDI higher than AHI also reflected RERA burden. Third,
HSATs were not performed. Typically, AHIs based on HSATs are less
than those based on in-laboratory polysomnography for 10 events/h
(19), emphasizing the need for
in-laboratory polysomnography. Fourth, no cardiovascular outcomes
were measured. The area under the ROC curve for BMI was 60.66%,
which indicated modest predictive ability. Finally, the study was
conducted in a single center with a sample size that could have led
to model instability. Additional large-sample studies with external
validation are required to confirm the present results.
In conclusion, patients with suspected OSA with BMI
≥26.03 kg/m2 may have higher RDI than AHI due to
moderate to severe OSA. These patients may require in-laboratory
polysomnography, whereas other patients may be tested by HSAT,
which may report comparable AHI values.
Acknowledgements
Not applicable.
Funding
Funding: No funding was received.
Availability of data and materials
The data generated in the present study may be
requested from the corresponding author.
Authors' contributions
NL, SK and KS conceived and designed the study. SS
and WB interpreted data. KS performed statistical analysis. All
authors have read and approved the final manuscript. NL, SK and KS
confirm the authenticity of all the raw data.
Ethics approval and consent to
participate
The present study protocol was approved (approval
no. HE641504) by the Ethics Committee in Human Research of Khon
Kaen University (Khon Kaen, Thailand) and all methods were
conducted following the principles of the Declaration of Helsinki.
The requirement for informed consent was waived due to the
retrospective nature of the study.
Patient consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing
interests.
References
|
1
|
Sönmez I, Vo Dupuy A, Yu KS, Cronin J, Yee
J and Azarbarzin A: Unmasking obstructive sleep apnea: Estimated
prevalence and impact in the United States. Respir Med.
248(108348)2025.PubMed/NCBI View Article : Google Scholar
|
|
2
|
Iannella G, Pace A, Bellizzi MG, Magliulo
G, Greco A, De Virgilio A, Croce E, Gioacchini FM, Re M, Costantino
A, et al: The global burden of obstructive sleep apnea. Diagnostics
(Basel). 15(1088)2025.PubMed/NCBI View Article : Google Scholar
|
|
3
|
Inami T, Seino Y, Otsuka T, Yamamoto M,
Kimata N, Murakami D, Takano M, Ohba T, Ibuki C and Mizuno K: Links
between sleep disordered breathing, coronary atherosclerotic
burden, and cardiac biomarkers in patients with stable coronary
artery disease. J Cardiol. 60:180–186. 2012.PubMed/NCBI View Article : Google Scholar
|
|
4
|
Rivas M, Ratra A and Nugent K: Obstructive
sleep apnea and its effects on cardiovascular diseases: A narrative
review. Anatol J Cardiol. 15:944–950. 2016.PubMed/NCBI View Article : Google Scholar
|
|
5
|
Kapur VK, Auckley DH, Chowdhuri S,
Kuhlmann DC, Mehra R, Ramar K and Harrod CG: Clinical practice
guideline for diagnostic testing for adult obstructive sleep apnea:
An American academy of sleep medicine clinical practice guideline.
J Clin Sleep Med. 13:479–504. 2017.PubMed/NCBI View Article : Google Scholar
|
|
6
|
Boulos MI, Kamra M, Colelli DR, Kirolos N,
Gladstone DJ, Boyle K, Sundaram A, Hopyan JJ, Swartz RH, Mamdani M,
et al: SLEAP SMART (Sleep apnea screening using mobile ambulatory
recorders after TIA/Stroke): A randomized controlled trial. Stroke.
53:710–718. 2022.PubMed/NCBI View Article : Google Scholar
|
|
7
|
Malhotra RK, Kirsch DB, Kristo DA, Olson
EJ, Aurora RN, Carden KA, Chervin RD, Martin JL, Ramar K, Rosen CL,
et al: Polysomnography for obstructive sleep apnea should include
Arousal-Based scoring: An American academy of sleep medicine
position statement. J Clin Sleep Med. 14:1245–1247. 2018.PubMed/NCBI View Article : Google Scholar
|
|
8
|
Tu X, Morgenthaler TI, Baughn J, Herold DL
and Lipford MC: Are scoring respiratory effort-related arousals
worth the effort? -A study comparing outcomes between 4% vs 3%
hypopnea scoring rules. Sleep Med. 124:396–403. 2024.PubMed/NCBI View Article : Google Scholar
|
|
9
|
Martinot JB, Le-Dong NN, Malhotra A and
Pépin JL: Respiratory effort during sleep and prevalent
hypertension in obstructive sleep apnoea. Eur Respir J.
61(2201486)2023.PubMed/NCBI View Article : Google Scholar
|
|
10
|
Knauert M, Naik S, Gillespie MB and Kryger
M: Clinical consequences and economic costs of untreated
obstructive sleep apnea syndrome. World J Otorhinolaryngol Head
Neck Surg. 1:17–27. 2015.PubMed/NCBI View Article : Google Scholar
|
|
11
|
Ogna A, Tobback N, Andries D, Tobback N,
Marques-Vidal P, Vollenweider P, Waeber G and Heinzer R: Prevalence
and clinical significance of respiratory Effort-related arousals in
the general population. J Clin Sleep Med. 14:1339–1345.
2018.PubMed/NCBI View Article : Google Scholar
|
|
12
|
Chung F, Abdullah HR and Liao P: STOP-Bang
questionnaire: A practical approach to screen for obstructive sleep
apnea. Chest. 149:631–638. 2016.PubMed/NCBI View Article : Google Scholar
|
|
13
|
Berry RB, Budhiraja R, Gottlieb DJ, Gozal
D, Iber C, Kapur VK, Marcus CL, Mehra R, Parthasarathy S, Quan SF,
et al: Rules for scoring respiratory events in sleep: Update of the
2007 AASM Manual for the Scoring of Sleep and Associated Events.
Deliberations of the Sleep Apnea Definitions Task Force of the
American Academy of Sleep Medicine. J Clin Sleep Med. 8:597–619.
2012.PubMed/NCBI View Article : Google Scholar
|
|
14
|
Sun GW, Shook TL and Kay GL: Inappropriate
use of bivariable analysis to screen risk factors for use in
multivariable analysis. J Clin Epidemiol. 49:907–916.
1996.PubMed/NCBI View Article : Google Scholar
|
|
15
|
Mickey RM and Greenland S: The impact of
confounder selection criteria on effect estimation. Am J Epidemiol.
129:125–137. 1989.PubMed/NCBI View Article : Google Scholar
|
|
16
|
Madley-Dowd P, Hughes R, Tilling K and
Heron J: The proportion of missing data should not be used to guide
decisions on multiple imputation. J Clin Epidemiol. 110:63–73.
2019.PubMed/NCBI View Article : Google Scholar
|
|
17
|
Kumar P, Rai DK and Kanwar MS: Comparison
of clinical and polysomnographic parameters between obese and
nonobese obstructive sleep apnea. J Family Med Prim Care.
9:4170–4173. 2020.PubMed/NCBI View Article : Google Scholar
|
|
18
|
Hsu WH, Yang CC, Tsai CY, Majumdar A, Lee
KY, Feng PH, Tseng CH, Chen KY, Kang JH, Lee HC, et al: Association
of low arousal threshold obstructive sleep apnea manifestations
with body fat and water distribution. Life (Basel).
13(1218)2023.PubMed/NCBI View Article : Google Scholar
|
|
19
|
Pordzik J, Seifen C, Ludwig K, Ruckes C,
Huppertz T, Matthias C and Gouveris H: Waiting for in-lab
polysomnography may unnecessarily prolong treatment start in
patients with moderate or severe OSA at home sleep apnea testing.
Nat Sci Sleep. 16:1881–1889. 2024.PubMed/NCBI View Article : Google Scholar
|
![Receiver operating characteristic
curve of body mass index on presence of respiratory disturbance
index higher than apnea-hypopnea index [area under the curve,
60.66% (95% confidence interval 46.53%, 74.78%)].](/article_images/br/24/5/br-24-05-02139-g00.jpg)
