Mesenteric fat thickness is associated with increased risk of obstructive sleep apnoea

Authors

  • Kin Hung Liu,

    Corresponding author
    1. Department of Imaging and Interventional Radiology, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    • Correspondence: Kin Hung Liu, Department of Imaging and Interventional Radiology, The Prince of Wales Hospital, Ngan Shing Street, Shatin, Hong Kong. Email: lkh605@ha.org.hk

    Search for more papers by this author
  • Winnie C.W. Chu,

    1. Department of Imaging and Interventional Radiology, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Kin Wang To,

    1. Department of Medicine and Therapeutics, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Fanny W.S. Ko,

    1. Department of Medicine and Therapeutics, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Susanna S.S. Ng,

    1. Department of Medicine and Therapeutics, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Jenny C.L. Ngai,

    1. Department of Medicine and Therapeutics, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Jeff Wai Sang Chan,

    1. Department of Imaging and Interventional Radiology, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Anil T. Ahuja,

    1. Department of Imaging and Interventional Radiology, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    Search for more papers by this author
  • David S.C. Hui

    1. Department of Medicine and Therapeutics, The Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong
    Search for more papers by this author

Abstract

Background and objective

Mesenteric fat is a type of intraperitoneal adipose tissue draining into portal circulation. The objective of this study was to investigate the relationships between mesenteric fat thickness and obstructive sleep apnoea (OSA) in patients with suspected OSA.

Methods

One hundred forty-nine subjects (men: 114; women: 35) with suspected OSA underwent ultrasound examinations of mesenteric, preperitoneal and subcutaneous fat thickness after overnight polysomnography. Body mass index (BMI) and neck circumference were recorded.

Results

The subjects with OSA (n = 130, apnoea/hypopnoea index (AHI) >5/h) had greater neck circumference, higher BMI, and greater mesenteric and preperitoneal fat thickness than those without OSA (n = 19, AHI ≤ 5/h). There was positive correlation of AHI with mesenteric (r = 0.43, P < 0.001) and preperitoneal fat thickness (r = 0.3, P < 0.001), whereas no significant association was observed between AHI and subcutaneous fat thickness (r = 0.09, P = 0.27). On multivariate logistic regression, after adjustments for gender, age, BMI, neck circumference, and preperitoneal and subcutaneous fat thickness, the mesenteric fat thickness had a positive association with the presence of moderate OSA and severe OSA, with odds ratios of 7.18 and 7.45 for every 1 cm increase in mesenteric fat thickness when AHI was defined as ≥15/h and AHI ≥ 30/h, respectively.

Conclusions

Mesenteric fat thickness is associated with increased risk of OSA, independent of other abdominal fat thickness, BMI and neck circumference. Sonographic measurement is potentially a useful tool for further evaluating the complex association of visceral fat, metabolic syndrome and OSA.

Abbreviations
AHI

apnoea/hypopnoea index

BMI

body mass index

CT

computed tomography

MRI

magnetic resonance imaging

OSA

obstructive sleep apnoea

PSG

polysomnography

Introduction

Obstructive sleep apnoea (OSA) is a cause of refractory hypertension, and is also associated with systemic inflammation, visceral fat deposition, insulin resistance and dyslipidemia.[1] Obesity, particularly central obesity, is regarded as the most important risk factor for OSA.[1, 2] The body fat location rather than the fat size is an important factor in assessing the risk of obesity-related disorders.[3] While it is found that obesity can lead to OSA through some mechanisms, recent studies have suggested that OSA may worsen obesity.[2] Therefore, a complex association between OSA and obesity exists, and studies investigating the regional fat distribution in relation to OSA severity would be helpful to evaluate this complex association. Several studies using computed tomography (CT) or magnetic resonance imaging (MRI) have shown a positive correlation of intra-abdominal fat accumulation with the risk of OSA.[4-6] Our group has previously reported that sonographic measurement of mesenteric fat thickness correlated better than visceral fat measured by MRI with cardiovascular risk factors,[7] in addition to being an independent determinant of fatty liver,[8] carotid intima-medial thickness[9] and metabolic syndrome.[10] Other researchers have also reported increased mesenteric fat in women with polycystic ovarian syndrome,[11] and in subjects with increased concentration of low-density lipoprotein particles and apolipoprotein A-II.[12] Mesenteric fat is a type of visceral fat draining into the portal circulation, and exhibits different metabolic characteristics from other abdominal fat deposits, such as subcutaneous and extraperitoneal fat.[13, 14] This may explain the closer association of the mesenteric fat thickness with various metabolic derangements in obese subjects. In addition, an animal study showed that the exposure to hypoxia resulted in the dysregulation of adiponectin production from the mesenteric fat tissues of obese mice.[15] This may indicate the importance of mesenteric fat in sleep apnoea subjects with visceral obesity.[15] To the best of our knowledge, there were no previous human studies examining mesenteric fat in OSA subjects. The objective of this study was to investigate the relationships between mesenteric fat thickness and the severity of OSA in a group of symptomatic patients suspected to have sleep-disordered breathing.

Methods

We recruited 149 Chinese subjects with snoring and/or other symptoms suggestive of OSA[16] who had been referred to a respiratory clinic. Subjects with diabetes mellitus and requiring treatment with metformin were excluded because metformin might affect the body fat distribution. Ethics approval was obtained from the joint ethics committee of the Prince of Wales Hospital and the Chinese University of Hong Kong. Informed written consent was obtained from all the subjects.

Sleep studies

Overnight polysomnography (PSG) was performed in all 149 subjects. The subjects with an apnoea/hypopnoea index (AHI) of more than five events per hour of sleep, as shown by overnight PSG (Healthdyne Alice 4, Atlanta, GA, USA), plus self–reported sleepiness, were defined to have OSA.[17, 18] The following data were collected during the overnight PSG: electrocardiogram, electrooculogram, electroencephalogram, submental electromyogram, bilateral anterior tibial electromyogram, chest and abdominal wall movement by inductance plethysmography, and airflow by a nasal pressure transducer (PTAF 2, Pro-Tech, Woodinville, WA, USA) with an oronasal thermistor and finger pulse oximetry.[17, 18] Sleep stages were scored according to standard criteria.[19] Apnoea was defined as cessation of airflow for ≥10 s, whereas hypopnoea was defined as at least 50% reduction in airflow for ≥10 s with an oxygen desaturation of >3% or an arousal. An arousal was scored when there was at least 3-s abrupt shift in electroencephalogram frequency to alpha or theta, or higher than 16 Hz, following ≥10 s of sleep, and when arising in rapid eye movement sleep, there must be a rise in electromyogram tone.[20] With the patient in upright position, neck circumference was measured at the level of the cricothyroid membrane.[21] Height and body weight were measured and recorded for the calculation of body mass index (BMI).

Ultrasound examination

All 149 subjects had ultrasound examination of the abdomen for the measurement of mesenteric, subcutaneous and preperitoneal fat thickness on the same day or 1 day after overnight PSG. The ultrasound examinations were performed with the ultrasound machine, Philips ATL HDI 5000 (Bothell, CA, USA). The sonographers performing the ultrasound examinations were blinded to the clinical and study results. A curvilinear transducer C5-2 MHz or C7-5 MHz was used for the measurement of mesenteric fat thickness, while curvilinear transducer C7-5 or linear array transducer L12-5 was used for the measurement of preperitoneal and subcutaneous fat thickness. The detailed methodology for the abdominal fat thickness measurement has been described in previous study.[7] The abdominal region, particularly the paraumbilical area, was scanned in each subject to identify the mesenteric leaves (Fig. 1). The mesenteric leaves were found to be tubular structures with linear echogenic peritoneal layers. The thin echogenic peritoneal layers separated different mesenteric leaves.[22] The thickness of mesenteric leaves was measured with measurement tool of the ultrasound machine. There were a variable number of mesenteric leaves, and not all of them could be visualized due to the partial obscuration by overlying bowel gas. Usually, six to ten mesenteric leaves were identified and measured. The measurements of the three thickest mesenteric leaves were taken, and the mean was used for the analysis. The preperitoneal fat thickness was measured in the longitudinal section, while the subcutaneous fat thickness was measured in the transverse section. The transducer was placed perpendicular to the skin surface. In the midline of the abdomen between the xiphoid process and the umbilicus, we measured the maximum preperitoneal and subcutaneous fat thicknesses three times, and the mean thickness value was used for analysis. The inter-operator reliability for measurement of mesenteric, preperitoneal and subcutaneous fat thickness was reported previously, with the intraclass correlation coefficients ranging from 0.89 to 0.97.[7]

Figure 1.

Ultrasonogram of mesenteric leaves. The bright reflective lines indicated by arrows are peritoneal surfaces, and the distance between two peritoneal surfaces is the thickness of mesentery (+……+). Inside the mesentery, a high level of echoes (open arrow) is detected in the centre, while a low level of echoes (arrowhead) is visualized in the periphery. These represent fat deposition in the mesentery.

Statistical analysis

The data are expressed as mean ± standard deviation. The mean values of the subjects with OSA (AHI > 5/h) and those without OSA (AHI ≤ 5/h) were compared by independent t tests. Pearson correlation was performed between the AHI and abdominal fat thickness, BMI and neck circumference. Analysis of variance was also performed to compare the differences of abdominal fat thickness, BMI and neck circumference among the apnoea subjects of different severity. Stepwise multiple regression was performed with the AHI (as continuous variable) being the dependent variable, and gender, BMI, age, neck circumference, and preperitoneal, subcutaneous and mesenteric fat thickness as the independent variables. Multivariate logistic regression was performed with the presence of varying degrees of OSA (as categorical variable) being the dependent variable in different statistical models, while the independent variables were the same as those in multiple linear regression. OSA was categorized as AHI ≥ 5/h, AHI ≥ 15/h and AHI ≥ 30/h in three different statistical models in the multivariate logistic regression.

Results

Table 1 shows the comparisons of the demographics, clinical parameters and the sonographic measurements between the subjects with and without OSA. No significant difference in age was observed between the two groups. The OSA subjects confirmed by PSG had greater neck circumference, higher BMI, and greater mesenteric and preperitoneal fat thicknesses than those without OSA.

Table 1. Comparison of clinical characteristics and ultrasound measurements of abdominal fat thickness between subjects with confirmed OSA and those without OSA
 Subjects with confirmed OSA (AHI > 5/h)Subjects without OSA (AHI ≤ 5/h)P-value
  1. AHI, apnoea/hypopnoea index; BMI, body mass index; OSA, obstructive sleep apnoea.
Number of subjects13019
Men : women103:2711:8
Age (years)50.5 ± 9.448.7 ± 10.80.45
BMI (kg/m2)28 ± 4.324.4 ± 3.2=0.001
Neck circumference (cm)39.3 ± 3.536.2 ± 3.1<0.001
Subcutaneous fat thickness (cm)2.58 ± 0.862.26 ± 0.680.13
Preperitoneal fat thickness (cm)1.6 ± 0.451.3 ± 0.390.006
Mesenteric fat thickness (cm)0.98 ± 0.290.73 ± 0.26=0.001
AHI (no./h)30.9 ± 23.33.2 ± 1.2<0.001
Mean SaO2 (%)92.8 ± 3.4995.5 ± 1.02<0.001
Arousal index (no./h)32.1 ± 18.223.1 ± 120.038

Association with AHI

There was positive correlation of AHI with mesenteric (r = 0.43, P < 0.001) (Fig. 2) and preperitoneal fat thicknesses (r = 0.3, P < 0.001), while no significant association was observed between AHI and subcutaneous fat thickness (r = 0.09, P = 0.27). The AHI also had significant correlation with neck circumference (r = 0.43, P < 0.001) and BMI (r = 0.44, P = < 0.001). The mesenteric and preperitoneal fat thicknesses, BMI, and neck circumferences increased significantly from mild to severe OSA (Table 2). In a stepwise multiple linear regression analysis, gender, mesenteric fat thickness and BMI were found to be significantly associated with the AHI (Table 3). In a multivariate logistic regression analysis (Table 4), all the independent variables did not have significant association with OSA when the dependent variable was defined as AHI ≥5/h. The mesenteric fat thickness had a positive and independent association with the presence of moderate and severe OSA (when stratified as AHI ≥ 15/h and AHI ≥ 30/h, respectively, in two statistical models) after adjustments for gender, age, BMI, neck circumference, and preperitoneal and subcutaneous fat thicknesses. The odds ratios of moderate OSA and severe OSA risk were 7.18 and 7.45 for every 1 cm increase in mesenteric fat thickness when the OSA was stratified as AHI ≥ 15/h and AHI ≥ 30/h, respectively. Gender was also found to be significantly associated with the presence of OSA in these two statistical models.

Figure 2.

Scatter plot of apnoea/hypopnoea index against mesenteric fat thickness (r = 0.43, P < 0.001).

Table 2. Comparison of abdominal fat thickness, BMI and neck circumference among subjects of different severity
 Without OSA (AHI ≤ 5/h) (n = 19)Mild OSA (5 < AHI < 15) (n = 39)Moderate OSA (15 ≤ AHI < 30) (n = 38)Severe OSA (AHI ≥ 30) (n = 53)P-value
  1. AHI, apnoea/hypopnoea index; BMI, body mass index; OSA, obstructive sleep apnoea.
Mesenteric fat thickness (cm)0.730.840.931.10<0.001
Preperitoneal fat thickness (cm)1.301.581.461.72=0.002
Subcutaneous fat thickness (cm)2.262.632.472.62=0.37
BMI (kg/m2)24.426.727.129.5<0.001
Neck circumference (cm)36.237.639.340.5<0.001
Table 3. Significant determinants of AHI on stepwise multiple regression analysis using gender, BMI, subcutaneous, and preperitoneal and mesenteric fat thicknesses as independent variables
 Regression coefficient (β)Standard error (SE)P-value
  1. The total variances explained by the variables in this statistical model are 0.32.
  2. AHI, apnoea/hypopnoea index; BMI, body mass index.
Gender15.24.01<0.001
Mesenteric fat thickness13.87.0=0.05
BMI1.950.47<0.001
Table 4. Multivariate logistic regression analysis
Model 1: OSA (AHI > 5/h) as dependent variable
 β ± SEP-valueOdds ratio (95% CI)
Gender1.03 ± 1.080.342.8 (0.34–23.1)
Age (years)0.034 ± 0.030.261.03 (0.98–1.1)
Preperitoneal fat thickness (cm)0.89 ± 0.870.302.43 (0.4513.3)
Subcutaneous fat thickness (cm)0.46 ± 0.600.451.58 (0.49–5.17)
Mesenteric fat thickness (cm)0.63 ± 1.370.651.88 (0.13–27.8)
BMI (kg/m2)0.15 ± 0.150.311.16 (0.87–1.55)
Neck circumference (cm)0.05 ± 0.170.751.06 (0.76–1.46)
Model 2: Moderate OSA (AHI ≥ 15/h) as dependent variable
 β ± SEP-valueOdds ratio (95% CI)
Gender2.5 ± 0.860.0312.2 (2.3–65.4)
Age (years)0.04 ± 0.0230.0781.04 (0.99–1.1)
Preperitoneal fat thickness (cm)−0.24 ± 0.550.660.79 (0.27–2.29)
Subcutaneous fat thickness (cm)0.37 ± 0.390.331.45 (0.68–3.1)
Mesenteric fat thickness (cm)1.97 ± 0.980.0457.18 (1.05–49.3)
BMI (kg/m2)0.10 ± 0.0990.301.11 (0.91–1.34)
Neck circumference (cm)0.002 ± 0.120.981.0 (0.8–1.3)
Model 3: Severe OSA (AHI ≥ 30/h) as dependent variable
 β + SEP-valueOdds ratio (95% CI)
  1. The dependent variable is the presence of OSA, while the independent variables include gender, age, preperitoneal, subcutaneous and mesenteric fat thickness, BMI, and neck circumference.
  2. AHI, apnoea/hypopnoea index; BMI, body mass index; CI, confidence interval; OSA, obstructive sleep apnoea; SE, standard error.
Gender2.32 ± 0.960.01410.3 (1.6–66)
Age (years)0.051 ± 0.0250.0371.05 (1.0–1.11)
Preperitoneal fat thickness (cm)0.52 ± 0.520.321.68 (0.60–4.67)
Subcutaneous fat thickness (cm)0.10 ± 0.340.761.11 (0.57–2.16)
Mesenteric fat thickness (cm)2.01 ± 0.970.0387.45 (1.12–49.6)
BMI (kg/m2)0.19 ± 0.0990.0611.21 (0.99–1.46)
Neck circumference (cm)−0.078 ± 0.120.50.93 (0.74–1.16)

Discussion

Increased visceral fat deposition was found to be important for assessing the risk of OSA in previous studies,[4-6] and the visceral fat amount was determined by cross-sectional imaging, such as MRI and CT. Based on a pilot study of 37 Chinese subjects, good correlation (r = 0.8) was found between mesenteric fat thickness measured by ultrasound and total visceral adiposity measured by MRI.[10] However, quantification of visceral fat on MRI and CT cannot reliably differentiate between intraperitoneal and extraperitoneal fat.[7] Both intra- and extraperitoneal fat exhibit different metabolic characteristics,[13, 14] and these may have different impact on the metabolic derangement, cardiovascular risks and the risk of OSA. Mesenteric fat is a type of intraperitoneal fat that drains into the portal vein (portal adipose tissue), whereas the preperitoneal and subcutaneous fat do not drain into the portal circulation. Higher lipolytic activity and lower response to anti-lipolytic effect of insulin are demonstrated in the portal than the non-portal adipose tissues. These lead to high rate of free fatty acid production[23] which further decreases the hepatic clearance of insulin. This results in insulin resistance and some other obesity-related disorders.[24] This might explain the previous study findings indicating that mesenteric fat thickness had better correlation with cardiovascular risk factors when compared with visceral fat determined on MRI.[7] It was also an independent determinant of fatty liver, carotid intima-media thickness and the metabolic syndrome.[8-10]

In this study, we have shown greater abdominal fat thickness, mesenteric fat thickness in particular, in subjects with OSA than those without OSA. There was also closer correlation of AHI with mesenteric fat thickness than with preperitoneal and subcutaneous fat thicknesses. The data suggest that the portal adipose tissue (mesenteric fat) might have different clinical impact or association when compared with the non-portal adipose tissues (preperitoneal and subcutaneous fat). In multivariate logistic regression, all independent variables, including mesenteric fat thickness, did not have significant association with OSA when the dependent variable was defined as AHI ≥5/h. This might be due to the small number of control subjects (19 subjects with AHI < 5/h). Neck circumference and BMI, known indexes of regional and general obesity respectively, also had significant correlation with AHI, as also demonstrated by previous studies.[4, 6] However, the effect of neck circumference and BMI were controlled in multivariate logistic regression analysis in which the mesenteric fat thickness was found to be independently associated with the risk of moderate OSA and severe OSA. The findings of this study suggested a stronger association of the intraperitoneal fat with increased risk of OSA with AHI ≥15/h.

In a study of 219 Japanese subjects with suspected OSA who had PSG performed, together with the measurement of neck circumference, total body fat and visceral fat levels assessed by bioelectrical impedance studies, neck circumference was associated with the OSA severity, independent of visceral obesity, particularly in non-obese patients.[25] This might be due to the different methods of visceral fat quantification between the current study and the Japanese study. In the current study, we used ultrasound to determine the mesenteric fat thickness, which specifically indicated the visceral fat amount draining into the portal circulation, while the Japanese study adopted the bioelectrical impedance method to estimate the overall visceral fat amount involving fat both draining or not draining into the portal circulation.

To our knowledge, there have been few studies investigating the association of mesenteric fat with OSA. In an ex-vivo study of obese mice, exposure of mesenteric fat tissue, but not subcutaneous tissue, to hypoxia resulted in suppression of adiponectin level in the culture media.[15] This may explain the nocturnal fall in circulating adiponectin concentrations seen in OSA patients with visceral obesity.[15] The exact pathogenesis of mesenteric fat in relation to the obesity-linked disorders and OSA is not fully understood. It has been postulated that visceral obesity and insulin resistance may progressively result in worsening of the metabolic syndrome manifestations and OSA. Progressive deterioration of OSA may result in increased accumulation of visceral fat and the metabolic syndrome by providing a stress stimulus and inducing nocturnal increase of hormones, such as cortisol and insulin, which enhance visceral fat accumulation, metabolic derangements and cardiovascular complications.[26] These factors may represent a vicious cycle that includes visceral fat accumulation, the metabolic syndrome and OSA. Others have also suggested that in subjects with visceral obesity, the mass load of visceral fat might cause an increase in the activities of the inspiratory muscles by restricting the diaphragmatic motion. This could then lead to more negative inspiratory airway pressure, and hence collapse of the upper airway.[4]

The small number of control subjects (AHI ≤ 5/h) was a significant limitation of the study and might influence outcomes. Another limitation was that in the multivariate logistic models using AHI ≥ 15/h and AHI ≥ 30/h, the ‘control’ groups did not only include non-OSA subjects, but also OSA subjects, with AHI ≤ 15/h and AHI ≤ 30/h, respectively. In addition, the treatment effect of continuous positive airway pressure on mesenteric fat thickness has not been examined in this study, whereas Chin et al. have previously shown in an uncontrolled study with CT that continuous positive airway pressure could reduce intra-abdominal visceral fat without change in body weight.[3] In a randomized controlled trial over 3 months with crossover design comparing therapeutic continuous positive airway pressure (n = 43) versus sham continuous positive airway pressure (n = 43) in patients with moderate to severe OSA, there were significant differences in the reduction of subcutaneous fat and visceral fat measured by CT, with improvement of other metabolic abnormalities in favour of therapeutic continuous positive airway pressure, but mesenteric fat was not measured.[27]

In conclusion, mesenteric fat thickness measured on ultrasound was independently associated with the risk of OSA after adjustments for other abdominal fat thicknesses, BMI and neck circumference. There is a complex association and interactive effects among visceral fat, the metabolic syndrome and OSA, and this requires further exploration with large-scale and in-depth studies. In view of better representation of intraperitoneal visceral fat, strong association with OSA and the metabolic syndrome and its easier access in comparisons with CT and MRI, sonographic measurement of mesenteric fat thickness may be a useful method to investigate these complex relationships.

Acknowledgement

We need to thank the nurses and research assistants of the Respiratory Medicine Division, The Chinese University of Hong Kong, for their assistance and coordination in the study.

Ancillary