The relationship between obstructive sleep apnea and insulin resistance in patients with type 2 diabetes mellitus: a cross-sectional observational study

Abstract

Background

Increasing evidence suggests obstructive sleep apnea (OSA) is an independent risk factor for type 2 diabetes. Insulin resistance is the primary mechanism in the early stage of type 2 diabetes. The aim of this study was to assess the prevalence of OSA among patients with type 2 diabetes mellitus and to examine the association between OSA severity and insulin resistance.

Methods

This study adopted a cross-sectional observational design. We enrolled patients hospitalized for poorly controlled type 2 diabetes mellitus from the Endocrine Department of Beijing Chao-Yang Hospital and recorded their demographic and clinical data in an electronic case report form. Polysomnography (PSG) monitoring was performed for all the subjects. According to the apnea hypopnea index (AHI), the subjects were classified into the following three groups: the control group, AHI <5 events per hour; the mild-OSA group, 5 events per hour ≤ AHI <15 events per hour; and the moderate-to-severe-OSA group, AHI ≥15 events per hour.

Results

A total of 96 patients with type 2 diabetes were enrolled in this study, among whom 78 (81.28%) were diagnosed with OSA. The participants were categorized into three groups: a control group of 18 patients without OSA, a mild-OSA group of 54 patients, and a moderate-to-severe OSA group of 24 patients. Analysis revealed that the C-peptide area under the curve (AUCcp) was significantly greater in the moderate-to-severe OSA group compared to the control group. Additionally, the homeostasis model assessment of insulin resistance (HOMA-IR) was significantly higher in both the moderate-to-severe OSA and mild OSA groups relative to the control group. Multiple stepwise regression analysis revealed that the AHI and AUCcp were positively correlated (r=0.323, P=0.001). Both the AHI and body mass index (BMI) were positively correlated with the HOMA-IR index (P=0.007 and 0.02, respectively).

Conclusions

A high prevalence of OSA was observed among patients hospitalized for poorly controlled type 2 diabetes; further, its severity was positively correlated with insulin resistance.

Highlight box.

Key findings

• The prevalence of obstructive sleep apnea (OSA) in hospitalized patients with type 2 diabetes is 81.25%. OSA is positively related to insulin resistance in patients with type 2 diabetes mellitus.

What is known and what is new?

• OSA is an independent risk factor for type 2 diabetes.

• The prevalence rate of OSA in patients with type 2 diabetes was high, and OSA and its severity were associated with insulin resistance in patients with type 2 diabetes.

What is the implication, and what should change now?

• Patients with type 2 diabetes, especially obese and snoring patients, should be alert to OSA, and sleep breathing monitoring should be carried out when necessary. Monitoring blood glucose is recommended for patients with OSA, particularly in those with severe, untreated disease.

Introduction

Obstructive sleep apnea (OSA) is a common disease in sleep-disordered breathing (SDB) patients. Research shows that due to the aging population and increasing obesity rate, approximately 33% of adults in the United States have OSA (1). Epidemiological surveys in several cities in China also show that the prevalence rate of OSA varies from 3% to 7% (2); However, the prevalence rate is likely to be underestimated due to the low disease awareness of patients with OSA (3). Diabetes is a common disease, research shows that approximately 34 million people in the United States have diabetes (4).

At present, it is believed that the main pathogenesis of type 2 diabetes is insulin resistance and defects in pancreatic islet B-cell function, and insulin resistance is the main mechanism involved in the early stage of the disease. OSA is characterized by recurrent episodes of apnea and/or hypopnea during sleep, often accompanied by snoring. These events result in intermittent hypoxia, hypercapnia, and sleep fragmentation, thereby triggering a cascade of effects including sympathetic activation, increased oxidative stress, and a systemic inflammatory response, etc. Studies have shown that sleep deprivation or sleep fragmentation can reduce insulin sensitivity, cause insulin resistance and impaired glucose tolerance, which are related to type 2 diabetes (5-8). Study indicate that sleep fragmentation is also associated with complications in type 2 diabetes (9). Consequently, sleep fragmentation in patients with OSA may contribute to diabetes complications and to diabetes itself. Intermittent hypoxia triggers a cascade of pathological conditions, including systemic inflammation, oxidative stress, sympathetic activation, mitochondrial dysfunction, and endothelial dysfunction. These disturbances collectively contribute to reduce insulin secretion, decreased cellular glucose uptake, and hyperglycemia (10-12). Increasing evidence shows that OSA is closely related to metabolic syndrome and type 2 diabetes (13,14). OSA may aggravate the severity of type 2 diabetes (15,16), and lead to poor blood sugar control in type 2 diabetes patients (17). Domestic research shows that the prevalence rate of OSA in patients with type 2 diabetes is approximately 60–67% (18,19). Despite the close relationship between OSA and type 2 diabetes, a large proportion of type 2 diabetes patients are still not adequately treated for OSA (20).

This cross-sectional study enrolled type 2 diabetes patients admitted to our hospital’s endocrinology department for suboptimal glycemic control. The aims were to determine the prevalence of OSA in this inpatient population and to explore the associations of OSA and its severity with insulin resistance and glycemic control. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2122/rc).

Methods

Subjects

This study included diabetic patients who were admitted to the Endocrinology Department of Beijing Chao-Yang Hospital of Capital Medical University from July 2011 to September 2016. The inclusion criteria for patients were as follows: a clear history of type 2 diabetes; a fasting blood glucose concentration ≥7.0 mmol/L (126 mg/dL) and/or a random blood glucose concentration ≥11.1 mmol/L (200 mg/dL) in patients with type 2 diabetes. The exclusion criteria were as follows: (I) left ventricular ejection fraction <40% and heart failure; (II) combined chronic obstructive pulmonary disease, pulmonary heart disease, or other respiratory diseases that can cause hypoxemia; (III) recent oxygen therapy or continuous positive airway pressure (CPAP) treatment; (IV) a history of stroke or severe consciousness disorders; (V) combined thyroid dysfunction, pituitary adenoma, or other endocrine disorders; (VI) severe liver or kidney dysfunction or malignant tumors; (VII) recent infection, surgery, or trauma due to stress conditions; and (VIII) unwillingness to participate or inability to complete all the examination items in this study due to severe illness. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study protocol was approved by the Ethics Committee of Beijing Chao-Yang Hospital (No. 2023-Ke-450) and informed consent was obtained from all the patients.

Data collection

Patient demographic data, including general information (including sex and age) and physical measurements such as height, weight, neck circumference, waist circumference, hip circumference, neck circumference, waist circumference, and hip circumference, were collected after admission, and body mass index (BMI) and the waist-hip ratio (WHR) were calculated. Medical history data, including diabetes duration, diabetes treatment, hypertension status, coronary heart disease status, hyperlipidemia status, snoring status, smoking status, drinking status and family history of diabetes, were collected. Laboratory tests included the oral glucose tolerance test (OGTT), glycated hemoglobin (HbA1c), blood lipids, and high-sensitivity C-reactive protein (hs-CRP). The C-peptide area under the curve (AUCcp) was calculated, which reflects the patient’s insulin secretion. According to the homeostasis model assessment (HOMA) model (21), the homeostasis model assessment of insulin resistance (HOMA-IR) was calculated to assess insulin resistance. The calculation method for the measured values is as follows:

$$\text{AUCcp}=0.5\times \left({\text{C}}_0+{\text{C}}_{180}\right)+{\text{C}}_{30}+{\text{C}}_{60}+{\text{C}}_{120}$$ [1]

$$\text{HOMA-IR}=\frac{{\text{G}}_0\times {\text{I}}_0}{22.5}$$ [2]

(Note: C₀, fasting C-peptide value; C₃₀, 30-minute C-peptide value; C₆₀, 60-minute C-peptide value; C₁₂₀, 120-minute C-peptide value; C₁₈₀, 180-minute C-peptide value; G₀, fasting blood glucose value; I₀, fasting insulin value).

All subjects underwent polysomnography (PSG) (Embla N7000, Embla Systems, Kanata, Canada) after their blood glucose levels had stabilized, and no treatment for OSA administered during the hospital stay. Subjects are prohibited from consuming alcohol, tea, coffee, or sedative medications, and from smoking on the day of the sleep monitoring. The sleep monitoring duration shall not be less than 7 hours. All data shall be automatically stored in a computer, with subsequent manual verification and correction performed the following day prior to review of the analyzed results by a senior physician at the Chaoyang Hospital Sleep Respiratory Center. Sleep monitoring scores were analyzed in accordance with the criteria recommended by the American Academy of Sleep Medicine. Apnea hypopnea index (AHI), lowest percutaneous oxygen saturation (LSpO₂), mean percutaneous oxygen saturation (MSpO₂), and percentage of time spent SpO₂ below 90% (SIT90%) were recorded. According to AHI, the subjects were divided into the three groups: control group: AHI <5 events/hour; mild group: 5 events/hour ≤ AHI <15 events/hour; moderate-to-severe group: AHI ≥15 events/hour (22). Sleep is divided into four stages: phase I sleep, phase II sleep, phase III sleep, and rapid eye movement (REM) sleep. The percentages of phase I+II sleep, phase III sleep, and REM sleep were calculated, respectively.

Statistical analysis

The SPSS 25.0 (IBM Corp., Armonk, NY, USA) statistical software was used for statistical analysis. The insulin values and HOMA-IR values at 5 time points were non normally distributed, and their natural logarithms were used to present a normal distribution. The baseline characteristics are reported as the medians (quartiles) or means ± standard deviations for continuous parameters and as frequencies (percentage) for categorical parameters. The comparison between multiple sets of econometric data that conform to normal distribution is conducted using one-way analysis of variance, and the comparison between two groups is conducted using least significant difference (LSD) method. Non normally distributed was used non parametric tests. The comparison between count data is conducted using Chi-squared test. Correlation analysis and multiple stepwise regression analysis were used to evaluate the correlations between variables with differences between groups and between the AUCcp and HOMA-IR. P<0.05 indicated a statistically significant difference.

Results

During the study period, 345 patients were admitted to the Endocrinology Department of Chaoyang Hospital, and 303 patients had diabetes. Overall, 96 patients, including 56 males and 40 females, were included in this study (see Figure 1 for the enrollment flow chart), with an average age of 51.44±10.74 years; 78 patients (81.28%) were diagnosed with OSA through PSG monitoring. The control group included 18 patients; the mild OSA group included 54 patients; and the moderate-to-severe OSA group included 24 patients.

Figure 1.

Flowchart of patient inclusion and exclusion. CPAP, continuous positive airway pressure.

The general characteristics and related blood biochemical indicators of each group of patients are shown in Table 1. There was no significant difference in sex, age, duration of diabetes, hypertension, coronary heart disease history, hyperlipidemia history, smoking history, or family history of diabetes between there groups. There was no significant difference in the WHR among the three groups. The BMI and neck circumference of the moderate-to-severe OSA group were significantly greater than those of the control group and the mild-OSA group. There was no statistically significant difference in total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglycerides, or apolipoprotein. A among the groups. The hs-CRP of the moderate-to-severe OSA group was significantly greater than that of the other two groups.

Table 1. Baseline patient characteristics.

Variables	Moderate-severe group (AHI ≥15/h) (n=24)	Mild group (5/h ≤ AHI <15/h) (n=54)	Control group (AHI <5/h) (n=18)	P value
Male	15 (62.50)	30 (55.56)	11 (61.11)	0.82
Age (years)	51.75±9.87	52.31±10.37	48.56±12.53	0.43
BMI (kg/m2)	28.99±4.12†‡	24.98±3.23	24.82±2.67	<0.001
Neck circumference (cm)	41.17±3.90†‡	37.94±3.78	37.44±4.07	0.002
WHR	0.93±0.05	0.92±0.06	0.89±0.05	0.07
Diabetes duration (years)	5.46±6.59	5.15±5.64	4.94±5.17	0.96
Untreated patients	7 (29.17)	19 (35.19)	8 (44.44)	0.59
Snoring	23 (95.83)†‡	36 (66.67)†	7 (38.89)	<0.001
Hypertension	13 (54.17)	18 (33.33)	5 (27.78)	0.14
Coronary heart disease	2 (8.33)	6 (11.11)	3 (16.67)	0.71
Hyperlipidemia	14 (58.33)	29 (53.70)	8 (44.44)	0.67
Smoking	10 (41.67)	19 (35.19)	5 (27.78)	0.65
Drinking	5 (20.83)	12 (22.22)	2 (11.11)	0.50
Family history of diabetes	15 (62.50)	24 (44.44)	11 (61.11)	0.24
CHOL (mmol/L	5.44±1.27	5.04±1.07	5.36±2.25	0.45
HDL-C (mmol/L)	1.19±0.32	1.24±0.35	1.19±0.31	0.69
LDL-C (mmol/L)	3.38±1.86	3.13±0.95	3.31±1.84	0.61
TG (mmol/L)	2.19 (2.29)	1.43 (1.14)	1.55 (2.1)	0.17
Lp (a) (mg/dL)	26.45 (23.75)	22.85 (12.12)	21.07 (18.53)	0.51
hs-CRP (mg/L)	2.90 (5.29)†‡	1.80 (2.49)	1.61 (1.78)	0.040

Data are presented as mean ± SD or n (%). ^†, significantly different from control group; ^‡, significantly different from mild group. AHI, apnea hypopnea index; BMI, body mass index; CHOL, total cholesterol; HDL-C, high-density lipoprotein cholesterol; hs-CRP, high-sensitivity C-reactive protein; LDL-C, low-density lipoprotein cholesterol; Lp (a), lipoprotein (a); SD, standard deviation; TG, triglyceride; WHR, waist-to-hip ratio.

Comparisons of blood glucose, insulin, C-peptide, the AUCcp, HbA1c, and the HOMA-IR index among the three groups of patients are shown in Table 2. The OGTT showed no statistically significant differences in blood glucose levels or HbA1c. The data for insulin concentrations at 30 minutes (I₃₀), 60 minutes (I₆₀), 120 minutes (I₁₂₀), and 180 minutes (I₁₈₀) after fasting and glucose intake were all not normally distributed. Therefore, the natural logarithm of these values was converted to a normal distribution before comparison. Except for the I₃₀, the insulin levels at the 5 time points were significantly greater in the moderate-to-severe OSA group than those in the control and mild-OSA groups. The C-peptide levels were significantly greater in the moderate-to-severe OSA group than those in the control group.

Table 2. Metabolic parameters.

Variables	Moderate-severe group (AHI ≥15/h) (n=24)	Mild group (5/h ≤ AHI <15/h) (n=54)	Control group (AHI <5/h) (n=18)	P value
Glucose (G0) (mmol/L)	7.88±2.02	8.23±3.13	8.32±2.85	0.85
30- minute glucose (G30) (mmol/L)	13.15±3.12	12.91±3.58	12.54±3.34	0.85
60- minute glucose (G60) (mmol/L)	16.87±3.42	17.36±4.47	15.92±3.66	0.43
120-minute glucose (G120) (mmol/L)	17.85±3.26	19.39±5.19	18.62±4.30	0.39
180- minute glucose (G180) (mmol/L)	13.88±3.67	15.14±5.65	15.27±4.02	0.54
LN (I0)	2.11±0.83†‡	1.65±0.89	1.28±0.39	0.005
LN (I30)	2.60±1.00†	2.18±0.92	1.95±1.04	0.08
LN (I60)	3.08±0.76†‡	2.54±1.02	2.36±0.52	0.02
LN (I120)	3.39±0.82†‡	2.78±1.01	2.42±0.73	0.003
LN (I180)	3.02±0.72†‡	2.50±0.96	2.29±0.51	0.01
C-peptide (C0) (ng/mL)	2.59±0.70†‡	2.08±0.85	1.79±0.60	0.003
30-minute C-peptide (C30) (ng/mL)	3.41±1.39†‡	2.72±1.14	2.27±0.81	0.006
60-minute C-peptide (C60) (ng/mL)	4.58±2.14†	3.76±1.64	2.90±1.09	0.008
120-minute C-peptide (C120) (ng/mL)	6.46±3.23†	5.45±2.88	4.12±2.02	0.04
180-minute C-peptide (C180) (ng/mL)	5.78±1.82†	5.05±2.24	4.00±1.91	0.03
AUCcp	18.64±7.42†	15.50±6.76	12.19±4.68	0.009
LN (HOMA-IR)	1.03±0.86†‡	0.56±0.94	0.23±0.58	0.01
HbA1c (%)	9.85±2.22	9.91±2.42	9.46±2.11	0.77

Data are presented as mean ± SD. ^†, significantly different from control group; ^‡, significantly different from mild group. AHI, apnea hypopnea index; AUCcp, C-peptide area under the curve; HbA1c, glycated hemoglobin; HOMA-IR, homeostasis model assessment of insulin resistance; LN, natural logarithm; I₀, insulin; I₃₀, 30-minute insulin; I₆₀, 60- minute insulin; I₁₂₀, 120-minute insulin; I₁₈₀, 180-minute insulin; SD, standard deviation.

Using the AUCcp to represent insulin secretion, there were significant differences among the control group, mild-OSA group, and moderate-to-severe OSA group, with the moderate-to-severe OSA group exhibiting significantly greater insulin secretion than the control group. The logarithm of HOMA-IR showed a normal distribution, and the LN(HOMA-IR) of OSA patients in the mild and moderate-to-severe groups was significantly greater than that in the control group (showed in Figure 2).

Figure 2.

AUCcp and HOMA-IR index in the three groups. (A) Comparison of AUCcp among the three groups of patients. (B) Comparison of LN(HOMA-IR) among the three groups of patients. AUCcp, C-peptide area under the curve; CI, confidence interval; HOMA-IR, homeostasis model assessment of insulin resistance; LN, natural logarithm.

Correlation analysis was conducted between BMI, neck circumference, AHI, LSpO₂, MSpO₂, SIT90%, apnea-hypopnea time/total sleep time (AHT/TST), longest apnea time (LAT), I+II sleep, III sleep, and REM sleep as a percentage of total sleep time and the AUCcp. The AHI was positively correlated with the AUCcp (r=0.330, P=0.001). A correlation analysis was conducted between the variables of interest and HOMA-IR, indicating that the AHI had the strongest correlation with HOMA-IR (r=0.400, P<0.001) (Figure 3).

Figure 3.

AUCcp and HOMA-IR correlation analysis. (A) The AHI was positively correlated with the AUCcp (r=0.330, P=0.001). (B) The AHI had the strongest correlation with HOMA-IR (r=0.400, P<0.001). AHI, apnea hypopnea index; AUCcp, C-peptide area under the curve; HOMA-IR, homeostasis model assessment of insulin resistance; LN, natural logarithm.

Discussion

In this study, PSG, OGTT and HbA1c levels were measured in hospitalized patients with type 2 diabetes, and diabetes was comprehensively and objectively assessed. The prevalence rate of OSA in patients with type 2 diabetes was high, and OSA and its severity were associated with insulin resistance in patients with type 2 diabetes.

Current research shows that the prevalence rate of OSA in patients with type 2 diabetes is greater than that in the general population (12,23,24). The American Sleep Heart Health Research conducted overnight PSG for all selected patients with type 2 diabetes and those without type 2 diabetes. The prevalence of a respiratory disorder index (RDI) ≥5 in patients with diabetes was 57.7%, while it was 42.6% in patients without diabetes (25). In this study, PSG was used to assess sleep respiratory events in patients with type 2 diabetes. The results showed that the percentages of 81.25% of patients with an AHI ≥5 and 55.21% of patients with an AHI ≥10 were greater than those reported in the American Sleep Heart Health Study.

In a study in South Korea, patients with type 2 diabetes underwent overnight PSG with an AHI ≥5 as the diagnostic criterion for OSA. The results showed that the prevalence rate of OSA in type 2 diabetes patients was 72% (26). These findings are similar to our research results and are slightly lower than those obtained in this study. The prevalence of OSA varies among different studies; one of the reasons may be the different diagnostic criteria used in the aforementioned studies, and differences in research subjects can also lead to inconsistent results.

In this study, hospitalized patients with type 2 diabetes were selected as the research objects. It was found that the levels of serum insulin and C-peptide were gradually increased in the control group, mild group and moderate to severe group. AUCcp can represent the amount of insulin secreted by B cells and is not affected by exogenous insulin. The AUCcp of the moderate to severe group is higher than that of the mild group and control group, indicating that there is significant hyperinsulinemia in the moderate to severe group at the same blood glucose level, indicating that the insulin resistance level of the moderate to severe group is more severe than that of the other two groups.

At present, the internationally recognized gold standard for measuring insulin resistance is the normal blood glucose insulin clamp technique, but this monitoring method is very expensive and time-consuming and is not conducive to large-scale implementation. This study calculated HOMA-IR as an indicator for evaluating insulin resistance by measuring fasting blood glucose and insulin levels. Studies have shown a strong correlation between HOMA-IR and insulin clamp technology (27,28), and HOMA-IR is currently a widely used indicator of insulin resistance. Our study revealed that the AHI was positively correlated with the HOMA-IR index (P=0.007). With the worsening of OSA, the degree of insulin resistance in patients becomes more serious, and insulin resistance is an important factor affecting the progression of type 2 diabetes. Insulin resistance may be the intermediate link between OSA and type 2 diabetes. Research on OSA and insulin resistance at home and abroad has focused mostly on individuals without diabetes, and it is speculated that insulin resistance may be an important cause of OSA-related abnormal glucose metabolism and type 2 diabetes (29). Multiple meta-analyses have suggested that CPAP can improve insulin resistance in OSA patients with type 2 diabetes (30,31).

Therefore, we conclude that the co-occurrence of type 2 diabetes and OSA is common clinically. Moreover, the severity of OSA shows a positive correlation with the level of insulin resistance in these patients, which likely underlies the observed challenge in managing their blood glucose levels. Increased severity of sleep apnea is associated with an increased risk of diabetes, and the risk may be partially explained by hypoxemia and arousal. In OSA, intermittent nocturnal hypoxia is a key stimulus for insulin resistance. It activates the sympathetic nervous system, thereby elevating stress hormones such as cortisol and catecholamines, which in turn promote the development of insulin resistance (32).

Despite its high prevalence of OSA in patients with type 2 diabetes, OSA is often underrecognized and undertreated. Patients with type 2 diabetes, especially obese and snoring patients, should be alert to OSA, and sleep breathing monitoring should be carried out when necessary (33). Monitoring blood glucose is recommended for patients with OSA, particularly in those with severe, untreated disease. Patients with moderate to severe OSA should be treated with continuous positive airway pressure. Follow-up sleep testing should be performed to assess the effectiveness of treatment.

There are limitations to this study. First, it is a single-center clinical study, and second, the study did not include patient follow-up to investigate whether CPAP treatment for OSA can improve blood glucose and insulin resistance. Prospective multicenter research is urgent for further exploring the impact of OSA on type 2 diabetes and its pathogenesis.

Conclusions

This study revealed that the prevalence of OSA in hospitalized patients with type 2 diabetes is high, especially in obese and snoring patients. OSA is positively related to insulin resistance in the patients with type 2 diabetes, which makes it difficult to control the blood glucose of type 2 diabetes patients. However, the diagnostic and treatment rates of OSA in patients with type 2 diabetes are remarkably low in clinical settings. Therefore, awareness of type 2 diabetes patients with OSA should be increased in special patients groups, PSG should be performed when necessary.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study protocol was approved by the Ethics Committee of Beijing Chao-Yang Hospital (No. 2023-Ke-450) and informed consent was obtained from all the patients.

Data Sharing Statement

Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2122/dss

jtd-18-02-152-dss.pdf ^{(168.1KB, pdf)}

DOI: 10.21037/jtd-2025-aw-2122