Research Article ISSN 2641-4333

Neurology - Research & Surgery

Correlating BMI, BP and Neck Circumference with AHI to predict OSA

Amira Ishag-Osman1, Brandon Barsky2,3, Andrew Dakkak2, Serena Spaleny2, Nadir Osman1,2 and Edith
Mensah-Osman1,2*
EENA Comprehensive Neurology & Sleep Center, Boynton
1

Beach FL. *
Correspondence:
Edith Mensah-Osman, M.D. PhD, EENA Comprehensive
2
Charles E. Schmidt College of Medicine-FAU, Boca Raton FL. Neurology & Sleep Center, Boynton Beach FL.

Nationwide Children’s Hospital/THE Ohio State University,

3
Received: 02 November 2020; Accepted: 19 November 2020
Columbus OH.

Citation: Ishag-Osman A, Barsky B, Dakkak A, et al. Correlating BMI, BP and Neck Circumference with AHI to predict OSA.
Neurol Res Surg. 2020; 3(1): 1-7.

ABSTRACT
Study Objective: Obstructive sleep apnea (OSA) is a cause of multiple significant comorbidities and affects
hundreds of millions of patients. The prevalence of OSA for adults between the ages of 30 and 70 years old is
estimated to be 26%, while the prevalence for the general adult population of more than 18yrs is as high as
38%. Even though OSA affects such a large proportion of the population, the majority of cases are undiagnosed,
and approximately 90% of US adults are untreated. These health factors and the economic impact highlight the
substantial importance of having accessible and accurate screening tools for OSA.

Subjects and Methods: This is a retrospective analysis of adult volunteer subjects who were recruited for this
study. Overall well-being was assessed via questionnaire survey, and physical attributes were extracted from EHR.
Daytime saliva and serum were collected from participants ranging between 11am-4pm. PSG was used for the
diagnosis of OSA according to AASM guidelines.

Results: BMI did not correlate with a high blood pressure (BP) in the newly diagnosed OSA patient pool, as well
as in the general pool of patients. BMI correlated better with neck circumference in the OSA patients. A high BMI
was associated with a high AHI value in the newly diagnosed OSA patients. Systolic and diastolic BP correlated
with levels of AHI but not for neck circumference in the OSA patients. ESS was not a good predictor of OSA in
patients. Levels of DHEA, afternoon cortisol and 17-OH progesterone in saliva of OSA patients were significantly
lower when compared to the normal reference range for each marker.

Conclusion: A composite value of large BMI and neck circumference in addition to a high systolic blood pressure
may be good indicators of AHI values to assess for OSA risk. Serum glucose was not a good predictor of OSA.
However, abnormal levels of products from the adrenal cortex in OSA patients may indicate dys-regulation of the
gluconeogenic pathway that may serve as early biomarkers easily accessible from the saliva.

Keywords the negative impact on population health, the American Academy

Obstructive Sleep Apnea, Body mass index, Apnea-hypopnea index. of Sleep Medicine (AASM) released an analysis estimating the
economic burden of undiagnosed OSA at approximately $150
Introduction billion a year. Additionally, an estimated annual savings of
The risks of hypertension, type II diabetes mellitus (T2DM), approximately $100 billion a year would be produced if all cases
hypercholesteremia, depression, gastro-esophageal reflux disease, of OSA in the US were diagnosed and treated [3]. These health and
heart disease, metabolic dysregulation, and stroke are tied to OSA economic factors highlight the substantial importance of having
and are increased when OSA is left undiagnosed [1,2]. In addition to easily accessible and accurate screening tools for OSA.
Neurol Res Surg, 2020 Volume 3 | Issue 1 | 1 of 7
Current accessible screening for OSA is largely reliant on the were diagnosed via attended polysomnography for confirmation
following: the STOP-BANG (SB) questionnaire, the Epworth of diagnosis within the AASM accredited center and as per routine
Sleepiness Scale (ESS), and the 4-Variable screening tool (4-V) clinical practice. All PSG data were scored by a certified registered
[4,5]. The SB questionnaire is more targeted to OSA and consists polysomnography technologist (RPSGT) using American
of eight yes/no questions. The first four questions relate to self- Academy of Sleep Medicine (AASM) criteria. Apnea–hypopnea
assessment, while the subsequent four questions relate to clinical index (AHI) was used in the diagnosis and assessment of severity
measurements. It has been found to have the highest sensitivity according to AASM approved guidelines [4].
compared to the other two, but importantly it also has the lowest
specificity [5]. The ESS is a non-specific measure of daytime Clinical evaluation included a sleep history, physical examination,
sleepiness. It consists of eight questions that are answered on a and follow up under the supervision of a board-certified sleep
scale from zero to three that reflect the likelihood of falling asleep medicine specialist. Questionnaires were completed and medical
during certain daytime activities. The ESS provides the second histories were obtained before the subjects underwent sleep studies.
highest specificity, however it is less specific for more severe OSA Daytime sleepiness was assessed by the Epworth Sleepiness Scale
and is prone to false-negatives [4,5]. The 4-V screening tool is an (ESS); ESS scores range from 0 to 24; a score >10 indicates
equation that uses gender, BMI, blood pressure, and self-reported excessive daytime sleepiness [9].
snoring. It equates to: OSA = (gender*4) + (BMI category value)
+ (BP category value) + (snoring*4). This equation provides the Overall well-being was assessed via a questionnaire survey at
highest specificity, only when values are >14, at which point the beginning and at follow-up during the study. Additionally,
sensitivity is only 51% [5]. Overall, each of these screening tools data from health records were collected, recorded, analyzed,
contains inherent weaknesses, and guidelines from the Journal of and compared for changes in blood pressure, body weight, neck
Clinical Sleep Medicine state none should be used diagnostically circumference, serum glucose and lipids, and saliva for mid-day
[4]. DHEA-S, cortisol, and 17-OH progesterone at initial diagnosis of
OSA. Data from all other patients not evaluated for OSA were
The gold standard objective diagnostic procedure for OSA is analyzed and used for comparison where relevant.
a polysomnographic recording [6]. Polysomnography (PSG)
is an overnight measurement of several physiological signals, Measurements
including electroencephalography (EEG) to measure sleep and a The questionnaire survey assessment tool was divided into three
comprehensive recording of respiratory parameters. The recording sections: socio-demographic factors, anthropometric measures,
is annotated by a sleep technician and evaluated, together with and biochemistry. The questionnaire gathered information on
symptomatology, to diagnose the presence of sleep-related demographics such as age, gender, and educational background;
disorders [7]. The number of annotated respiratory events divided risk factors of chronic diseases such as smoking, alcohol intake,
by the sleep time yields the apnea-hypopnea index (AHI) [8]. AHI diet, and physical exercise; prevalence of chronic diseases
thresholds of 5, 15, 30 events/h correspond to mild, moderate and including hypertension, diabetes mellitus, and dyslipidemia.
severe OSA, respectively.
The anthropometric measures evaluated height and body weight as
While polysomnography is a preferred tool for diagnosing OSA, it measured in the upright position to the nearest 0.5cm and 0.1kg,
has its challenges [9], and is not as cost-effective or as convenient respectively. Volunteers underwent measurement of neck or waist
as a biomarker. OSA biomarkers provide a promising avenue of circumference taken at specified times throughout the study.
research that could overcome the flaws of the above screening and
diagnostic tools by cost-effectively providing high specificity and Blood pressure for each subject was taken in the sitting position
sensitivity while also measuring disease severity. after 30 minutes of rest and recorded in their medical chart at each
visit. Three readings each of systolic and diastolic blood pressures
The objective of this pilot study was to determine if a combination were recorded with an interval of five minutes at the least, and the
of vital signs and biomarkers from the saliva may predict OSA mean of each measure was used for the data analysis.
and its severity in adult patients. Such combinations may be
further investigated and validated for measurable biochemical Sample collection and biomarker testing
characteristics that can be associated with the severity of OSA, as Saliva, collected during the day between 11 am and 4 pm with
a diagnostic tool that is both more cost-effective and convenient no specific prior restrictions to subjects were immediately stored
than polysomnography. in -80 C and then shipped with coded labels to a third party
commercial laboratory for processing.
Material and Methods
The study was approved by the institutional review board at Biochemical analysis
participating center and carried out in accordance with The Code Sampling of easily accessible bodily fluids (serum and urine) were
of Ethics of the Declaration of Helsinki. Written informed consent collected and sent under appropriate standardized conditions to an
was obtained from all subjects. Adult volunteer OSA patients external commercial laboratory for biochemical analysis.

Neurol Res Surg, 2020 Volume 3 | Issue 1 | 2 of 7

Results circumference (C). In the subset of patients newly diagnosed with
Figure 1 OSA we observe trends that correlate systolic and diastolic blood
Approximately 247 participants were enrolled and pooled for this pressures with AHI (A). Additionally, the peaks for BMI and AHI
study. To determine if there is a relationship between systolic and overlap and coincide with higher BMI correlating with increased
diastolic blood pressure (BP) relative to neck circumference in the AHI values (B). However, no significant association or trend is
newly diagnosed OSA patients, systolic and diastolic values were observed for AHI and neck circumference (C).
plotted against neck circumference from data obtained from EMR
charts. Figure 1a shows no specific pattern of correlation of BP Figure 3
with neck circumference in the general pool of patients. Figure 1b Figure 3 shows no difference in the pattern of expression or
shows a linear trend towards an increase in neck circumference trend for the Epworth Sleepiness Scale (ESS) between the newly
with higher systolic and diastolic BP levels in newly diagnosed diagnosed OSA patients and the rest of the participants.
OSA patients. Patients with a primary diagnosis of OSA had a
BMI greater than 27.53; however, the BMI of OSA patients did Figure 4
not correlate with BP Figure 1d. The general pool demonstrated a DHEA-S and 17-OH progesterone are commercially available
trend that higher systolic and diastolic BP could be associated with tests that can be easily performed from saliva of patients. Here we
higher BMI levels (Figure 1c). measured the values of DHEA-S and 17-OH in afternoon saliva in
OSA patients, which were significantly lower than the established
Figure 2 reference range. Additionally, while measuring afternoon saliva
In this study, OSA was assessed based on the Apnea-hypopnea cortisol may be challenging due to several factors that may
index (AHI) which were plotted as arbitrary units to compare influence its secretion, our data shows a trend towards a lower than
with systolic and diastolic blood pressure (A), BMI (B) and neck reference value in saliva of afternoon cortisol levels.

Figure 1: Blood Pressure (BP) correlates better with neck circumference compared to BMI in newly diagnosed OSA patients. No specific pattern of
correlation of BP with neck circumference or BMI can be observed in the general pool of patients.
A) Separation of systolic (blue) and diastolic (yellow) BP values in arbitrary units, plotted against neck circumference from a general pool of patients.
B) Separation of systolic (blue) and diastolic (yellow) BP values in arbitrary units, plotted against neck circumference from newly diagnosed patients
with OSA.
C) Separation of systolic (blue) and diastolic (yellow) BP values in arbitrary units, plotted against body mass index (BMI) from a general pool of
patients.
D) Separation of systolic (blue) and diastolic (yellow) BP values in arbitrary units, plotted against body mass index (BMI) from newly diagnosed
patients with OSA.

Neurol Res Surg, 2020 Volume 3 | Issue 1 | 3 of 7

Figure 2: Values for AHI and BMI plotted at arbitrary units showed correlations for both newly diagnosed patients with OSA as well as all subjects.
Overall, the BMI for the newly diagnosed OSA patients trended higher compared to all subjects.
Values for BMI (red) and AHI (blue) were plotted for both newly diagnosed OSA patients and all subjects. Insert shows the levels in all subjects. Red horizontal
line is the cut-off value point for all subjects at ~ 100 arbitrary units*. The highest BMI peak was at ~ 140** arbitrary units for the OSA patients.

Figure 3: No difference in the Epworth sleepiness scale (EPSS) obtained from questionnaire in new diagnosed OSA patients and the general pool of patients.
The score values of EPSS obtained from the general pool of patients (blue) compared to patients newly diagnosed with OSA (red).
Neurol Res Surg, 2020 Volume 3 | Issue 1 | 4 of 7
Figure 4: Set at a normal reference range of 100% the relative values of morning (11am) DHEA, Cortisol and 17-OH progesterone plotted from the
saliva of newly diagnosed OSA patients show significantly lower levels compared to reference range.
Reference range set at a 100% the graphs shows 11am saliva expression of DHEA-s(blue), cortisol (red) and 17-OH progesterone (green).

Figure 5: Serum levels of triglycerides (TG), HDL and Glucose in mg/dl measured in patients Figure 5: Newly diagnosed with OSA showed higher
levels of glucose and triglycerides compared to all subjects. There were no observed differences with HDL levels.
Levels of triglycerides (TG) in red, Glucose (green) and HDL (blue) were assessed in serum of patients newly diagnosed with OSA. Insert shows the
levels in all subjects. Red horizontal line is the cut-off value point for all subjects at ~350 mg/dl*. The highest peak was for glucose in the OSA patients
at ~ 700mg/dl ** followed by triglycerides at ~500mg/dl; compared to 350mg/dl and 250mg/dl, respectively for all patients.

Figure 5 lab results. The values for glucose and TG trended higher in the
OSA has been directly tied to metabolic dysfunction, specifically, subjects diagnosed with OSA and peaked at twice the levels for
glucose regulation and T2DM. Studies are suggestive of OSA glucose in the OSA subjects as compared to the overall subjects.
contributing to impaired glucose metabolism secondary to sleep
fragmentation, sympathetic excitation, and intermittent hypoxia Discussion
effecting pancreatic B-cell function, insulin sensitivity, and Serious public health consequences are associated with being
systemic inflammation. We plotted the levels of triglycerides (TG), overweight and obese [10]. Obesity is a major risk factor for
HDL, and glucose obtained from the EMR of subject’s routine the development of OSA. OSA is an established risk factor for
Neurol Res Surg, 2020 Volume 3 | Issue 1 | 5 of 7
insulin resistance and other cardio-metabolic disorders [11]. The OSA with a high AHI score at diagnosis. These anthropometric
enigma remains whether OSA has any causal role in the adverse measures which are routinely obtained during doctor’s visit, in
metabolic profile, independent of or beyond that due to obesity. addition to commonly used sleep assessment tools, can help easily
Approximately one third of the US population was overweight identify potential OSA patients that can then be fully evaluated
or obese in 2003-2004 and since then, percentages of overweight and diagnosed with polysomnographic (PSG) testing.
individuals have increased exponentially [12]. OSA is one of the
most serious medical squeals from obesity, in association with heart References
disease, hypertension and type 2 diabetes [13]. The costs associated 1. Pinto JA, Ribeiro DK, Cavallini AF, et al. Comorbidities
with diagnosis of OSA remain very high and cumbersome i.e. with Associated with Obstructive Sleep Apnea: a Retrospective
PSG [14,15]; thus, there is a need for identifying biomarkers that Study. Int Arch Otorhinolaryngol. 2016; 20: 145-150.
are easily accessible for early identification and stratification of
2. Kent BD, McNicholas WT, Ryan S. Insulin resistance, glucose
OSA at point-of-care. Biomarkers associated with OSA may also
intolerance and diabetes mellitus in obstructive sleep apnoea.
be used as pharmaceutical targets that reduce weight and mitigate
J Thorac Dis. 2015; 7: 1343-1357.
adverse effects of obesity, including hyperlipidemia and insulin
resistance. 3. https://aasm.org/economic-burden-of-undiagnosed-sleep-
apnea-in-u-s-is-nearly-150b-per-year/
This study sought to compare trends of biometric measurements in 4. https://aasm.org/resources/clinicalguidelines/diagnostic-
OSA patients versus the general patient population and to discern testing-osa.pdf
whether specific combination of metrics could predict OSA, as 5. h t t p : / / w w w . a c c . o r g / l a t e s t - i n - c a r d i o l o g y /
well as serve as an objective biomarker to further the evaluation articles/2015/07/14/11/04/screeing-tools-for-the-obstructive-
for OSA in a general patient population. sleep-apnea-for-the-cardiovascular-clinician
6. Rundo JV, Downey R 3rd. Polysomnography. Handb Clin
Neck circumference has been suggested to be a more reliable Neurol. 2019; 160: 381-392.
measure of obesity [16], and a risk factor for hypertension [17], in
relation to increasing systolic and diastolic blood pressures. There 7. Jafari B, Mohsenin V. Polysomnography. Clin Chest Med.
are several conflicting reports on the studies previously done on 2010; 31: 287-297.
the association of neck circumference with AHI or OSA [18-20]. 8. Borsini E, Nogueira F, Nigro C. Apnea-hypopnea index in
In our study, while a correlation was not observed with AHI, a sleep studies and the risk of over-simplification. Sleep Sci.
large neck circumference was associated with a greater BMI. 2018; 11: 45-48.
9. Johns MW. A new method for measuring daytime sleepiness:
As previously reported by others, ESS was not effective in the Epworth sleepiness scale. Sleep. 1991; 14: 540-545.
differentiating and discriminating newly diagnosed OSA patients 10. Mesarwi OA, Sharma EV, Jun JC, et al. Metabolic dysfunction
with the general population [21,22]. in obstructive sleep apnea: A critical examination of underlying
mechanisms. Sleep Biol Rhythms. 2015; 13: 2-17.
DHEA-S [23,24] and 17-OH progesterone [25] are commercially
11. Lam DC, Lam KS, Ip MS. Obstructive sleep apnoea, insulin
available tests that are performed from saliva of patients for
resistance and adipocytokines. Clin Endocrinol (Oxf). 2015;
routine clinical assessment and evaluation. This study showed that
82: 165-177.
the values of DHEA-S and 17-OH Progesterone were significantly
lower than the established reference range in the newly diagnosed 12. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of
OSA patients. This finding warrants further investigation and with overweight and obesity in the United States, 1999-2004.
a larger population pool for validation. JAMA. 2006; 295: 1549-1555.
13. Gami AS, Caples SM, Somers VK. Obesity and obstructive
Additionally, while measuring afternoon saliva cortisol may be sleep apnea. Endocrinol Metab Clin North Am. 2003; 32: 869-
challenging due to several factors that may influence its secretion 894.
and collection [26], our data shows a trend towards a decrease 14. Bravata DM, Lightner N, Yaggi HK, et al. Economic
in afternoon salivary cortisol levels. This also warrants further Assessment of 4 Approaches to the Diagnosis and Initial
investigation and validation in a larger pool of newly diagnosed Treatment of Sleep Apnea. Respir Care. 2018; 63: 50-61.
OSA patients. Furthermore, the effect of treatment with CPAP
15. Toraldo DM, Passali D, Sanna A, et al. Cost-effectiveness
on these parameters can determine if the combination of these
strategies in OSAS management: a short review. Acta
biomarkers can predict the diagnosis of OSA [27,28].
Otorhinolaryngol Ital. 2017; 37: 447-453.
In summary, our data suggest the following combination: High 16. Hingorjo MR, Qureshi MA, Mehdi A. Neck circumference
BP=high BMI=High AHI, where high BMI correlates with a as a useful marker of obesity: a comparison with body mass
larger neck circumference. A combination of high systolic and index and waist circumference. J Pak Med Assoc. 2012; 62:
diastolic blood pressures, large BMI, and abnormally low DHEA 36-40.
levels may be an objective biomarker signal for the presence of 17. Zhou JY, Ge H, Zhu MF, et al. Neck circumference as an
Neurol Res Surg, 2020 Volume 3 | Issue 1 | 6 of 7
 independent predictive contributor to cardio-metabolic 2018; 9: 1828.
syndrome. Cardiovasc Diabetol. 2013; 12: 76. 24. Whetzel CA, Klein LC. Measuring DHEA-S in saliva: time of
18. Tom C, Roy B, Vig R, et al. Correlations between Waist day differences and positive correlations between two different
and Neck Circumferences and Obstructive Sleep Apnea types of collection methods. BMC Res Notes. 2010; 3: 204.
Characteristics. Sleep Vigil. 2018; 2: 111-118.
25. Mylonas PG, Makri M, Georgopoulos NA, et al. Adequacy
19. Ahbab S, Ataoğlu HE, Tuna M, et al. Neck circumference, metabolic of saliva 17-hydroxyprogesterone determination using various
syndrome and obstructive sleep apnea syndrome; evaluation of
collection methods. Steroids. 2006; 71: 273-276.
possible linkage. Med Sci Monit. 2013; 19: 111-117.
26. Halpern CT, Whitsel EA, Wagner B, et al. Challenges of
20. Davidson TM, Patel MR. Waist circumference and sleep
disordered breathing. Laryngoscope. 2008; 118: 339-347. measuring diurnal cortisol concentrations in a large population-
based field study. Psychoneuroendocrinology. 2012; 37: 499-
21. Baiardi S, La Morgia C, Sciamanna L, et al. Is the Epworth
Sleepiness Scale a useful tool for screening excessive daytime 508.
sleepiness in commercial drivers? Accid Anal Prev. 2018; 110: 27. Weinstock TG, Wang X, Rueschman M, et al. A controlled
187-189. trial of CPAP therapy on metabolic control in individuals with
22. Trimmel K, Żebrowska M, Böck M, et al. Wanted: a better impaired glucose tolerance and sleep apnea. Sleep. 2012; 35:
cut-off value for the Epworth Sleepiness Scale. Wien Klin 617-625.
Wochenschr. 2018; 130: 349-355. 28. Semelka M, Wilson J, Floyd R. Diagnosis and Treatment of
23. Francavilla VC, Vitale F, Ciaccio M, et al. Use of Saliva Obstructive Sleep Apnea in Adults. Am Fam Physician. 2016;
in Alternative to Serum Sampling to Monitor Biomarkers 94: 355-360.
Modifications in Professional Soccer Players. Front Physiol.

© 2020 Ishag-Osman A, et al. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License

Neurol Res Surg, 2020 Volume 3 | Issue 1 | 7 of 7