Abstract
Purpose:
Insomnia is frequently co-morbid with obstructive sleep apnea (OSA); the effect of insomnia or co-morbid insomnia and OSA (OSA+I) on associated metabolic outcomes in adults with type 2 diabetes (T2D) remains unclear. This study in adults with T2D compared metabolic outcomes among persons with OSA, insomnia, or OSA+I.
Methods:
This study analyzed baseline data from the Diabetes Sleep Treatment Trial of persons recruited for symptoms of OSA or poor sleep quality. Home sleep studies determined OSA presence and severity. Insomnia was evaluated using the Insomnia Severity Index. Height and weight to calculate body mass index (BMI) and blood for laboratory values were obtained. Multivariate general linear models were used to examine the impact of the type of sleep disorder and sociodemographic, lifestyle, and sleep risk factors on metabolic outcomes.
Results:
Participants (N=253) were middle-aged (56.3 ± 10.5 years), White (60.5%), obese (mean BMI of 35.3 ± 7.1 kg/m2), and male (51.4%) with poor glucose control (mean HbA1c of 8.0 ± 1.8 %). Most participants had OSA+I (42.7%) or insomnia only (41.0%). HbA1c and BMI differed among the sleep disorder groups. Additionally, in the adjusted models, having insomnia only, compared to OSA only, was associated on average with higher HbA1c levels (b = 1.08 ± .40, p <0.007) and lower BMI (b = −7.03 ± 1.43, p <0.001).
Conclusions:
Findings suggest that insomnia frequently co-exists with OSA, is independently associated with metabolic outcomes in adults with T2D, and should be considered in investigations of the effects of OSA in persons with T2D.
Keywords: diabetes, obstructive sleep apnea, insomnia, metabolic risk factors
Introduction
Insomnia and obstructive sleep apnea (OSA) are common among adults with type 2 diabetes (T2D) with an estimated prevalence of 25% and 50%−70%, respectively.1–4 Insomnia is characterized by difficulties falling asleep, difficulties with maintaining sleep including frequent awakenings or the inability to return to sleep after awakenings, and/or unwanted early awakenings resulting in significant distress or functional impairments.5 OSA is characterized by intermittent apneas and hypopneas during sleep caused by upper airway obstruction.6 Insomnia and OSA are independent risk factors for metabolic diseases. Both disorders result in inflammation, oxidative stress, impaired glucose tolerance, and insulin resistance.7–11
There is emerging evidence that insomnia is frequently co-morbid with OSA;12 however, the differences in demographic, behavioral and sleep risk factors, and associated metabolic outcomes between adults with T2D and OSA, insomnia, or co-morbid insomnia and OSA (OSA+I) remains unclear and understudied. A greater understanding of the characteristics of adults with these sleep disorders and their metabolic consequences may direct better T2D management and treatment. Therefore, among a sample of adults with T2D and either OSA, insomnia, or OSA+I, this study first examined the univariate associations between these three sleep disorder groups and sociodemographic characteristics, behavioral and sleep risk factors, and metabolic outcomes. Secondly, the study examined the associations between sleep disorder group and cardiometabolic outcomes while adjusting for sociodemographic and risk factors.
Research Design and Methods
This study was a secondary analysis of baseline data from a multi-site randomized sham-controlled trial (Diabetes Sleep Treatment Trial [DSTT], R01-DK096028) of persons with T2D recruited for symptoms of OSA (snoring/breath holding) or “poor sleep quality.” The primary purpose of the DSTT was to determine whether treatment of OSA with continuous positive airway pressure (CPAP) would result in improved glucose control and diabetes self-management behaviors compared to participants randomized to sham-CPAP.13 Screening criteria for the parent study were self-report of T2D, CPAP naïve, ambulatory, and not pregnant. To ensure the safety of the individuals randomized to the sham-CPAP group, not being employed in a safety sensitive occupation or history of an automobile accident or near miss within the last 12 months was required.13 Participants meeting the screening criteria were scheduled for an assessment where informed consent was obtained prior to further evaluation. Participants were eligible to be included in this secondary analysis if there were valid data that indicated they had either OSA, insomnia, or comorbid OSA+I. The study was approved by the University of Pittsburgh’s Human Research Protection Office and listed in ClinicalTrials.gov (NCT01901055).
Measures
Evaluation of Sleep
OSA presence and severity were determined by home ApneaLink Plus® sleep studies The ApneaLink Plus® includes a nasal cannula, pulse oximeter with finger pulse sensory, and respiratory effort sensor to record respiratory flow, breathing sounds, blood-oxygen saturation, pulse, and respiratory effort. All sleep studies were verified by a board-certified sleep physician. Studies were considered valid if they had either 2-hours of data with an apnea + hypopnea index (AHI) ≥10 or 4-hours of data. Sleep studies with inadequate sleep duration or poor study quality were repeated. While an AHI of > 5 is considered mild OSA, a conservative approach of using an AHI ≥10 to consider a participant as having OSA was used in the study.
Insomnia presence and severity was determined by the Insomnia Severity Index (ISI), a 7-item questionnaire that asks on a 5-point Likert scale from “0” (no problem) to “5” (very severe problem) regarding the frequency, severity, and impact of insomnia symptoms.14,15 The ISI is considered a reliable and valid measure with a score ≥10 having an 86% sensitivity and 88% specificity for detecting insomnia in community-dwelling adults. An ISI ≥10 was the cut point used in this secondary analysis to classify participants as having insomnia.
Perceived daytime sleepiness was measured using the Epworth Sleepiness Scale (ESS).16 The ESS is a validated measure to evaluates subjective daytime sleepiness by querying the likelihood of falling asleep in 8 situations, ranging from lying down after lunch (without alcohol) to sitting in traffic. Potential scores for each item range from “0” (no likelihood) to “3” (a strong likelihood). Total ESS scores > 10 have a 93.5% sensitivity and a 100% specificity to detect excessive daytime sleepiness.
Perceived sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI).17 The PSQI is a 19-item validated measure of different sleep components (duration, efficiency, subjective quality, disturbances) with a Cronbach’s alpha of .83 and a correlation coefficient of .85. Global PSQI scores greater than 5 are indicative of poor sleep.
Average sleep duration was determined by a 7-day sleep diary where participants indicated their time in bed, estimated sleep latency, wake after sleep onset, and wake time.
Clinical Evaluation
Participants were measured wearing light clothing without shoes for height and weight to calculate body mass index (BMI, kg/m2). Venipunctures were performed to obtain laboratory values for glycated hemoglobin (HbA1c), total cholesterol, low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C). Participant were provided a BodyMedia Armband® accelerometer to wear for one week to measure physical activity, specifically daily step count. This device has been validated by Jakicic and colleagues18 and used in a previous pilot study.19
Demographic and Health Questionnaires
Questionnaires collected information on participant age, sex, race, education, and perceived financial difficulty. Participants were asked how many years since they were diagnosed with diabetes and whether they were on insulin.
The self-reported data (ISI, ESS, PSQI, demographic questionnaire, and health questionnaire) were completed at home and returned by mail. Sleep diary data were collected during the seven days after the in-person assessment. All data were collected prior to randomization for the DSTT.
Statistical Analysis
Analyses were conducted using SAS version 9.4 (SAS, Cary, NC, USA). The data were screened for missingness, outliers, and normality. The significance level was set at 0.05 for two-sided hypothesis testing. Descriptive statistics for continuous variables were reported as mean ± standard deviation (SD) and categorical variables reported as frequency count and percentage (%). Univariate general linear models (GLM) were conducted to examine the differences in demographics, behavioral and sleep risk factors, and cardiometabolic outcomes among the three sleep disorder groups (OSA, insomnia, and OSA+I). Post-hoc analyses with the Tukey adjustment for multiple group comparison were conducted to examine the pairwise differences between the three groups.
Next, multivariate GLM was used to examine associations between sleep disorder group and the set of cardiometabolic outcomes simultaneously while adjusting for demographics and risk factors. Variables with a p-value of <. 20 based on the univariate analyses were included in the models. Additionally, the multicollinear relationships between the sleep risk factors were examined. Sleep duration was highly correlated with sleep quality (r = −0.57, P=<0.001) where the lower the PSQI score the better the sleep quality. As a result, the predictor variables included in the models were age, sex, marital status, race, financial distress, insulin use, mean daily steps, daytime sleepiness, and sleep quality.
Results
Sociodemographics Characteristics
Participants (N=253) were middle-aged (56.3 ± 10.5 years), white (60.5%), obese (BMI of 35.3 ± 7.1 kg/m2), and male (51.4%) with fair to poor glucose control (HbA1c of 8.0 ± 1.8%). Most participants had OSA+I (42.7%) or insomnia only (41.1%), while 16.2% had OSA only. There were statistically significant differences in age, sex, marital status, race, and perceived financial difficulty among the three groups (OSA, insomnia, and OSA+I) with p-values < .005 (Table 1).
Table 1.
Sociodemographic characteristics of total sample and comparison by sleep disorder group
| Variable | Total Sample (N = 253) |
OSA (n = 41) |
Insomnia (n = 104) |
OSA+I (n = 108) |
p-value |
|---|---|---|---|---|---|
| Age, years | 56.3 ± 10.6 | 60.0 ± 10 | 53.4 ± 10.1a | 57.7 ± 10.5 | <0.001 |
| Sex | |||||
| Female | 123 (48.6%) | 11 (26.8%) | 67 (64.4%)a | 45 (41.7%) | <.001 |
| Male | 130 (51.4%) | 30 (73.2%) | 37 (35.6%) | 63 (58.3%) | |
| Marital status | |||||
| Single | 144 (56.9%) | 22 (53.7%) | 73 (70.2%)b | 49 (45.4%) | 0.001 |
| Married/Partnered | 109 (43.1%) | 19 (46.3%) | 31 (29.8%) | 59 (54.6%) | |
| Race | |||||
| Non-White | 100 (39.5%) | 12 (29.3%) | 55 (52.9%)a | 33 (30.6%) | 0.001 |
| White | 153 (60.5%) | 29 (70.7%) | 49 (47.1%) | 75 (69.4%) | |
| Education | |||||
| <2-year degree | 116 (45.8%) | 17 (41.5%) | 47 (45.2%) | 52 (48.1%) | 0.754 |
| ≥2-year degree | 137 (54.2%) | 24 (58.5%) | 57 (54.8%) | 56 (51.9%) | |
| Financial difficulty | |||||
| “Some” to “extreme” | 99 (39.1%) | 8 (19.5%) | 55 (53.4%)b | 36 (33.3%) | <0.001 |
| “None” | 153 (60.5%) | 33 (80.5%) | 48 (46.6%) | 72 (66.7%) | |
| Diabetes duration, | |||||
| years | 11.16 ± 9.44 | 11.89 ± 11.16 | 10.21 ± 7.16 | 11.82 ± 10.67 | 0.409 |
| Insulin use | |||||
| Yes | 139 (54.9%) | 24 (58.5%) | 63 (60.6%) | 52 (48.1%) | 0.169 |
| No | 114 (45.1%) | 17 (41.5%) | 41 (39.4%) | 56 (51.9%) |
Data presented are mean ± standard deviation or number and percentage
OSA, obstructive sleep apnea; OSA+I, obstructive sleep apnea and insomnia.
Significant differences in age, sex, and race between the “insomnia only” group and “OSA only” and “OSA+I” groups.
Significant differences marital status and financial difficulty between the “insomnia only” and “OSA only” group
The adults with OSA and OSA+I were older than the adults with insomnia only (p < 0.001 and p = 0.003, respectively; Table 1). Participants with OSA and OSA+I were mostly male (p < 0.001), whereas females made up the largest proportion of those with insomnia only (p = 0.001). Participants with OSA+I were more likely to be married, while those with insomnia only were more likely to be single (p < 0.001). Those with OSA and OSA+I were mostly white (p = 0.01); while non-white individuals made up the largest proportion of those with insomnia only (p = 0.001). Lastly, individuals with OSA+I were less likely to report financial difficulty compared to individuals with insomnia only (p = 0.003).
Behavioral and Sleep Risk Factors
In terms of behavioral and sleep risk factors, there were statistically significant differences in sleep duration, daytime sleepiness, and sleep quality among the three groups (OSA, insomnia, and OSA+I) with p-values < .005 (Table 2). Adults with OSA reported longer mean sleep durations than the adults with OSA+I (p = 0.003) and insomnia only (p = 0.006). Similarly, those with OSA only reported less daytime sleepiness than the adults with OSA+I (p < 0.001) and insomnia only (p < 0.001). Likewise, the participants with OSA reported better sleep quality than the participants with OSA+I (p < 0.001) and insomnia only (p < 0.001) and the participants with OSA+I reported better sleep quality than the participants with insomnia only (p = 0.001). However, the mean PSQI score for all three groups was > 5 which indicates poor sleep quality. Although there were no statistically significant differences in mean daily steps among the three groups, the individuals with insomnia took more daily steps than the individuals with OSA+I (p < 0.04).
Table 2.
Comparison of risk factors for impaired metabolic outcomes in adults with type 2 diabetes and sleep disorders
| Variable | Total Sample | OSA | Insomnia | OSA+I | p-value |
|---|---|---|---|---|---|
| Daily Stepsa | 4692 ± 2847 | 4329 ± 2160 | 5195 ± 3347 | 4354 ± 2490 | 0.071 |
| Sleep Duration, hoursb | 6.0 ± 1.6 | 6.7± 1.5c | 5.7 ± 1.5 | 5.9 ± 1.6 | 0.002 |
| Daytime sleepiness | 10.4 ± 4.7 | 7.5 ± 4.1c | 10.7 ± 4.4 | 11.1 ± 4.8 | < 0.001 |
| Sleep quality | 10.4 ± 4.0 | 5.8 ± 2.2c | 12.1 ± 3.5d | 10.5 ± 3.7 | < 0.001 |
Data presented are mean ± standard deviation
OSA, obstructive sleep apnea; OSA+I, obstructive sleep apnea and insomnia.
Average of 7 days of BodyMedia Armband data
Self-reported sleep duration from sleep diary
Significant differences in sleep duration, daytime sleepiness, and sleep quality between the “OSA only” group and “insomnia” and “OSA+I” groups.
Significant differences in sleep quality between the “insomnia only” group and the “OSA+I” group
Metabolic Outcomes
There were statistically significant differences in total cholesterol, LDL-C, HbA1c, and BMI among the three groups (OSA, insomnia, and OSA+I) with p-values ranging from 0.02 to < 0.001 (Table 3). Participants with OSA and OSA+I had lower total cholesterol levels compared to the participants with insomnia only (p = 0.02 and p = 0.005, respectively). Similarly, LDL-C levels were lower in the OSA and OSA+I groups than in the insomnia only group (p = 0.04 and p = 0.007, respectively). Likewise, HbA1c levels were lower in the OSA and OSA+I groups compared to the insomnia only group (p = 0.01 and p = 0.05, respectively). However, BMI was higher in the OSA and OSA+I groups than the insomnia only group (p < 0.001 and p < 0.001, respectively).
Table 3.
Comparison of metabolic outcomes in adults with type 2 diabetes and OSA, insomnia, or OSA and insomnia
| Variable | Total Sample | OSA | Insomnia | OSA+I | p-value |
|---|---|---|---|---|---|
| Total cholesterol, mmol/L | 4.41 ± 1.06 | 4.15 ± 1.17 | 4.65 ± 1.08a | 4.25 ± 0.99 | 0.007 |
| HDL-C, mg/dL | 1.14 ± 0.32 | 1.09 ± 0.29 | 1.20 ± 0.37a | 1.11 ± 0.28 | 0.055 |
| LDL-C, mg/dL | 2.30 ± 0.91 | 2.13 ± 1.00 | 2.49 ± 0.93a | 2.16 ± 0.84 | 0.020 |
| HbA1c, % | 8.0 ±1.8 | 7.4 ± 1.4 | 8.4 ± 2.3a | 7.9 ± 1.3 | 0.008 |
| BMI, kg/m2 | 35.3 ± 7.1 | 37.9 ± 7.8 | 32.8 ± 6.0b | 36.7 ± 7.1 | <.001 |
Data presented are mean ± standard deviation.
OSA, obstructive sleep apnea; OSA+I, obstructive sleep apnea and insomnia; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; HbA1c, hemoglobin A1c; BMI, body mass index.
Levels of total cholesterol, HDL cholesterol, LDL cholesterol, HbA1c were significantly higher among “Insomnia only” participants than “OSA only” and “OSA+I” groups.
BMI was significantly higher among in “OSA only” and “OSA+I” groups than in “insomnia only” group.
Factors Associated with Worse Metabolic Outcomes
Based on the multivariate GLM models, total cholesterol, HDL-C, and LDL-C levels were not different based on sleep disorder group after adjusting for demographic and risk factor variables (Table 4). Being female was associated with higher total cholesterol level, while insulin use was associated with lower total cholesterol levels. Older age, being female, and more steps per day were associated in higher HDL cholesterol levels in the multivariate model. Only being female was associated with higher LDL cholesterol levels.
Table 4.
Multivariate generalized linear models on the associations of OSA, insomnia, and OSA and insomnia with metabolic outcomes
| Variable | Total cholesterol | HDL-C | LDL-C | ||||||
|---|---|---|---|---|---|---|---|---|---|
| b | SE | p-value | b | SE | p-value | b | SE | p-value | |
| OSA only (reference) | |||||||||
| Insomnia only | 13.20 | 8.87 | 0.138 | 0.05 | 2.66 | 0.986 | 8.20 | 7.99 | 0.306 |
| OSA+I | 6.32 | 8.09 | 0.436 | −1.01 | 2.43 | 0.679 | 1.09 | 7.39 | 0.883 |
| Age, years | −0.45 | 0.26 | 0.085 | 0.17 | 0.08 | 0.030 | −0.29 | 0.24 | 0.222 |
| Sex | |||||||||
| Female (Male reference) | 12.98 | 5.28 | 0.015 | 8.57 | 1.58 | <.001 | 9.86 | 4.77 | 0.040 |
| Marital status | |||||||||
| Partnered (Single reference) | −6.73 | 5.56 | 0.227 | 1.43 | 1.67 | 0.393 | −5.12 | 5.02 | 0.309 |
| Race | |||||||||
| Non-White (White reference) | 0.02 | 5.94 | 0.997 | 3.09 | 1.78 | 0.084 | 8.79 | 5.31 | 0.099 |
| Financial Distress | |||||||||
| Yes (No reference) | −8.46 | 5.65 | 0.136 | −0.83 | 1.69 | 0.627 | −5.33 | 5.07 | 0.294 |
| Insulin use | |||||||||
| Yes (No reference) | −12.67 | 5.03 | 0.013 | −0.26 | 1.51 | 0.865 | −4.15 | 4.94 | 0.402 |
| Daily steps | 0.001 | 0.001 | 0.270 | 0.001 | 0.0003 | 0.047 | 0.0005 | 0.0008 | 0.526 |
| Daytime sleepiness | 0.35 | 0.57 | 0.542 | 0.07 | 0.17 | 0.696 | 0.59 | 0.51 | 0.251 |
| Sleep quality | −0.10 | 0.75 | 0.894 | 0.26 | 0.22 | 0.241 | −0.46 | 0.69 | 0.489 |
| Variable | HbA1c | BMI | ||||
|---|---|---|---|---|---|---|
| b | SE | p-value | B | SE | p-value | |
| OSA only (reference) | a | b | ||||
| Insomnia only | 1.08 | 0.40 | <0.007 | −7.03 | 1.43 | <0.001 |
| OSA+I | 0.57 | 0.36 | 0.122 | −1.89 | 1.31 | 0.150 |
| Age, years | −0.02 | 0.01 | 0.150 | −0.27 | 0.04 | <0.001 |
| Sex | ||||||
| Female (Male reference) | −0.12 | 0.23 | 0.602 | 1.16 | 0.85 | 0.171 |
| Marital status | ||||||
| Partnered (Single reference) | 0.003 | 0.25 | 0.990 | −0.12 | 0.90 | 0.893 |
| Race | ||||||
| Non-White (White reference) | 0.32 | 0.26 | 0.224 | 0.48 | 0.95 | 0.617 |
| Financial Distress | ||||||
| Yes (No reference) | 0.11 | 0.25 | 0.674 | 0.60 | 0.90 | 0.504 |
| Insulin use | ||||||
| Yes (No reference) | 0.67 | 0.22 | 0.003 | −0.61 | 0.81 | 0.447 |
| Daily steps | 0.0001 | 0.0004 | 0.023 | −0.001 | 0.0002 | <0.001 |
| Daytime sleepiness | 0.02 | 0.03 | 0.370 | −0.002 | 0.09 | 0.979 |
| Sleep quality | −0.04 | 0.03 | 0.187 | −0.02 | 0.12 | 0.865 |
OSA, obstructive sleep apnea; OSA+I, obstructive sleep apnea and insomnia; HDL-C, high-density lipoprotein cholesterol; HbA1c, hemoglobin A1c, BMI, body mass index.
Type III SS test F-value = 4.03; p-value = 0.019
Type III SS test F-value = 18.59; p-value < 0.001
For both HbA1c and BMI there were statistically significant differences among the groups after adjusting for the demographic and risk factors variable (F-value = 4.03; p value = 0.019 and F-value = 18.59; p value < 0.001, respectively). Insomnia only, compared to OSA only, insulin use, and greater daily steps were associated with higher HbA1c levels. Based on the multivariate model, predictors for a lower BMI included insomnia only, compared to OSA only, older age, and greater daily steps.
Discussion
To the best of our knowledge, this is the first study to examine the association between behavioral and sleep risk factors and cardiometabolic outcomes in adults with T2D and sleep comorbidities (OSA, insomnia, or OSA+I). In our descriptive analyses, we identified sociodemographic differences in cardiometabolic risk factors among the groups. Individuals with insomnia only were younger and more likely to be female, single, non-white, and experiencing financial difficulties compared to those with OSA and OSA+I. Insomnia is more prevalent in women than in men,20,21 thus it is not surprising in the present study to find higher rates of women in both the insomnia and OSA-I group. The participants with OSA only had longer sleep duration, less daytime sleepiness, and better sleep quality compared to the participants in the insomnia only and OSA+I groups. Individuals with insomnia had worse metabolic outcomes including higher total cholesterol, LDL-C, and HbA1c levels while also having a higher HDL-C and lower BMI compared to those with OSA and OSA+I. Based on the multivariate models that include sociodemographic variables and lifestyle and sleep risk factors, there were statistically significant differences in HbA1c and BMI among the sleep disorder groups. Additionally, the insomnia only group was associated with higher HbA1c levels but a lower BMI compared to the OSA only group.
Knutson and colleagues (2011) reported associations between sleep disturbances and fasting glucose levels in middle-aged adults with diabetes. In the models adjusted for age, race, sex, BMI, education, and income, a 10% increase in sleep fragmentation was associated with a 9% increase in glucose, while insomnia was associated with a 23% increase.22 Frequent snoring symptoms, assessed using the Berlin Sleep Apnea questionnaire, was not associated with increased glucose levels.22 These results are consistent with our finding of higher HbA1c levels in the insomnia only group compared the OSA only group.
Among a sample of 1311 individuals with insomnia, Hein and colleagues found that approximately 14% had comorbid moderate-to-serve OSA defined by an AHI ≥15.23 This was less than the 43% with OSA+I from our study. However, we used a lower threshold of an AHI ≥ 10 for the diagnosis of OSA, which may explain the differences in prevalence. Similar to our study, they found that those with insomnia only were younger, more likely to be female, and had a lower BMI compare to those with OSA+I.23
From a community-based participatory research study of individuals without preexisting cardiovascular disease, Luyster and colleagues (2014) also found that individuals with insomnia only were more likely to be female compared to those with OSA or OSA+I.24 While their study did not compare the metabolic risk factors between the three sleep disorder groups, they did compare the individual sleep disorders groups to a control group without OSA or insomnia. Those in the OSA and OSA+I groups were more likely to have a BMI ≥ 30 than the those in the control group (odds ratio [OR] = 39.14, 95% confidence interval [CI] = 18.81 – 81.49 and OR = 25.50, 95% CI = 11.80 −55.11, respectively) after adjusting for age, sex, and race. Whereas those with insomnia were no more likely to have a BMI ≥ 30 than the control group.24 None of the three sleep disorder groups were at a higher risk of dyslipidemia than the control group.24 We also found a similar pattern in the insomnia group were they also had a lower BMI than the OSA or OSA-I group. Although all the participants in our study had T2D, the results from Luyster and colleagues (2014) suggests that this pattern may not be particular to T2D.24
Cho and colleagues (2018) examined OSA and OSA-I in a sample of Korean adults. The prevalence of diabetes and hyperlipidemia did not differ between the two groups.25 These results should be understood in the context of the limitations that were inherent in the study. For instance, dichotomous self-reported data on type 2 diabetes and hyperlipidemia were used by Cho and colleagues.25 In our analyses, we used objective measures that included HbA1c, total cholesterol, and LDL-C to inform our observation that individuals with insomnia only had worse cardiometabolic outcomes compared to those with OSA and OSA+I. Consistent with our study, Cho and colleagues reported a higher ratio of females in the OSA-I group compared to the OSA group (35.3% vs. 19.6%, p < .001).25 Greater daytime sleepiness (8.97 ± 5.03 vs. 7.84 ± 4.43, p = .011) and shorter sleep duration in minutes were reported by the participants with OSA+I compared to those with OSA only (313.06 ± 80.88 vs. 337.14 ± 70.28; p = .006).24 We also found greater daytime sleepiness in the OSA-I group suggesting that the dual burden of both sleep disorders worsens daytime function.
Strengths and Limitations
The study has several limitations. All participants had T2D; as a result, the findings may not be generalizable to all adults with OSA, insomnia, or OSA+I without comorbid T2D or the associated metabolic consequences. Blood pressure was not measured as part of the study. Thus, the study was unable to examine differences in blood pressure among the three groups or how blood pressure impacted the metabolic outcomes. Likewise, this study does not include the objective measures of sleep duration, sleep latency, or wake after sleep onset that could be obtained with a device such as actigraphy. The use of home sleep studies had potential positive and negative effects on the study. While sleeping at home is more ecologically valid, especially for persons with insomnia, the ApneaLink Plus® does not obtain sleep stages and therefore, studies may have not captured adequate sleep duration to determine whether individuals with insomnia had a high AHI when asleep.
A strength of this study was the examination of two sleep disorders, OSA and insomnia, to tease out the independent and joint contribution of each on metabolic outcomes in persons with T2D. The diverse sample with almost 50% women and 40% non-white persons is important to racial inclusivity in sleep research and may provide additional strength for generalizability. However, the parent study was not designed to evaluate the effect of insomnia in persons with T2D and the cross-sectional design prohibits inferring causality.
Summary
In conclusion, these findings add evidence that insomnia is highly prevalent in adults with T2D, and frequently coexists with OSA. In our sample, differences in age, sex, marital status, race, and perceived financial difficulty among the three sleep disorder groups (OSA, insomnia, and OSA+I) were present. Individuals with insomnia were younger and more likely to be female and non-white compared to those with OSA and OSA+I. Additionally, individuals with insomnia were more likely to be single and report financial difficulty compare to individuals with OSA+I. In regards to sleep risk factors, the individuals with OSA reported longer sleep duration, less daytime sleepiness, and better sleep quality than the individuals with insomnia and OSA+I. While, statistically significant differences in total cholesterol, LDL-C, HbA1c, and BMI among the three groups were present, after adjusting for sociodemographic, behavioral, and sleep factors, insomnia was independently associated with worse HbA1c and lower BMI compared to persons with OSA. Future investigations of OSA and metabolic outcomes should consider the possible presence and effect of insomnia individuals with T2D.
Funding:
R01-DK096028 (Chasens); UL1-TR001857, UL1-TR000005
Footnotes
Trial name: Diabetes-Obstructive Sleep Apnea Treatment Trial (NCT01901055), https://clinicaltrials.gov/ct2/show/NCT01901055; Registration date: July 17, 2013.
Conflicts of interest/Competing interests: ApneaLink Plus® devices were obtained by a loan agreement from ResMed, Inc. Continuous positive airway pressure (CPAP) and sham-CPAP devices obtained by a loan agreement from Philips-Respironics, Inc. The authors have no additional conflicts of interest to declare. Dr. Strollo has received consultancy fees and honoraria from Inspire Medical Systems, ResMed, Philips Respironics, Emmi Solutions, Jazz Pharmaceuticals, and Itamar and has a provisional patent for positive airway pressure with integrated oxygen.
Ethical approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the University of Pittsburgh’s Human Research Protection Office.
Informed consent: Informed consent was obtained from all individual participants included in the study.
References
- 1.Brooks B, Cistulli PA, Borkman M, et al. (1994) Obstructive sleep apnea in obese noninsulin-dependent diabetic patients: effect of continuous positive airway pressure treatment on insulin responsiveness. J Clin Endocrinol Metab 79:1681–1685. doi: 10.1210/jcem.79.6.7989475 [DOI] [PubMed] [Google Scholar]
- 2.Ford ES, Cunningham TJ, Giles WH, Croft JB (2015) Trends in insomnia and excessive daytime sleepiness among U.S. adults from 2002 to 2012. Sleep Med 16:372–378. doi: 10.1016/j.sleep.2014.12.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Foster GD, Sanders MH, Millman R, et al. (2009) Obstructive sleep apnea among obese patients with type 2 diabetes. Diabetes Care 32:1017–1019. doi: 10.2337/dc08-1776 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Punjabi NM, Sorkin JD, Katzel LI, Goldberg AP, Schwartz AR, Smith PL (2002) Sleep-disordered breathing and insulin resistance in middle-aged and overweight men. Am J Respir Crit Care Med 165:677–682. doi: 10.1164/ajrccm.165.5.2104087 [DOI] [PubMed] [Google Scholar]
- 5.American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders, 5th edn. American Psychiatric Publishing, Arlington, Virginia [Google Scholar]
- 6.Drager LF, Togeiro SM, Polotsky VY, et al. (2013) Obstructive sleep apnea: a cardiometabolic risk in obesity and the metabolic syndrome. J Am Coll Cardiol 62:569–576. doi: 10.1016/j.jacc.2013.05.045 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tasali E, Ip MS (2008) Obstructive sleep apnea and metabolic syndrome: alterations in glucose metabolism and inflammation. Proc Am Thorac 5:207–217. doi: 10.1513/pats.200708-139MG [DOI] [PubMed] [Google Scholar]
- 8.Lévy P, Bonsignore MR, Eckel J (2009) Sleep, sleep-disordered breathing and metabolic consequences. Eur Respir J 34:243–260. doi: 10.1183/09031936.00166808 [DOI] [PubMed] [Google Scholar]
- 9.Depner CM, Stothard ER, Wright KP Jr (2014) Metabolic consequences of sleep and circadian disorders. Curr Diab Rep 14:507. doi: 10.1007/s11892-014-0507-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Javaheri S, Redline S (2017) Insomnia and risk of cardiovascular disease. Chest 152:435–444. doi: 10.1016/j.chest.2017.01.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bigini EG, Chasens ER, Conley YP, Imes CC (2019) DNA methylation changes and improved sleep quality in adults with obstructive sleep apnea and diabetes. BMJ Open Diabetes Res Care 7:e000707. doi: 10.1136/bmjdrc-2019-000707 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ye L, Pien GW, Ratcliffe SJ, et al. (2014) The different clinical faces of obstructive sleep apnoea: a cluster analysis. Eur Respir J 44:1600–1607. doi: 10.1183/09031936.00032314 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chasens ER, Atwood CW, Burke LE, et al. (2019) Diabetes sleep treatment trial: premise, design, and methodology. Contemp Clin Trials 76:104–111. doi: 10.1016/j.cct.2018.11.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bastien CH, Vallières A, Morin CM (2001) Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Med 2:297–307. doi: 10.1016/s1389-9457(00)00065-4 [DOI] [PubMed] [Google Scholar]
- 15.Morin CM, Belleville G, Bélanger L, Ivers H (2011) The Insomnia Severity Index: psychometric indicators to detect insomnia cases and evaluate treatment response. Sleep 34:601–608. doi: 10.1093/sleep/34.5.601 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Johns MW (1991) A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 14:540–545. doi: 10.1093/sleep/14.6.540 [DOI] [PubMed] [Google Scholar]
- 17.Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ (1989) The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 28:193–213. doi: 10.1016/0165-1781(89)90047-4 [DOI] [PubMed] [Google Scholar]
- 18.Jakicic JM, Marcus M, Gallagher KI, et al. (2004) Evaluation of the SenseWear Pro Armband to assess energy expenditure during exercise. Med Sci Sports Exerc 36:897–904. doi: 10.1249/01.mss.0000126805.32659.43 [DOI] [PubMed] [Google Scholar]
- 19.Chasens ER, Korytkowski M, Sereika SM, Burke LE, Drumheller OJ, Strollo PJ Jr (2014) Improving activity in adults with diabetes and coexisting obstructive sleep apnea. West J Nurs Res 36:294–311. doi: 10.1177/0193945913500567 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zhang B, Wing YK (2006). Sex differences in insomnia: a meta-analysis. Sleep 29:85–93. doi: 10.1093/sleep/29.1.85 [DOI] [PubMed] [Google Scholar]
- 21.Hall MH, Kline CE, Nowakowski S (2015) Insomnia and sleep apnea in midlife women: prevalence and consequences to health and functioning. F1000Prime Rep 7:63. doi: 10.12703/P7-63 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Knutson KL, Van Cauter E, Zee P, Liu K, Lauderdale DS (2011) Cross-sectional associations between measures of sleep and markers of glucose metabolism among subjects with and without diabetes: the Coronary Artery Risk Development in Young Adults (CARDIA) Sleep Study. Diabetes Care 34:1171–1176. doi: 10.2337/dc10-1962 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hein M, Lanquart JP, Loas G, Hubain P, Linkowski P (2017) Prevalence and risk factors of moderate to severe obstructive sleep apnea syndrome in insomnia sufferers: a study on 1311 subjects. Respir Res 18:135. doi: 10.1186/s12931-017-0616-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Luyster FS, Kip KE, Buysse DJ, Aiyer AN, Reis SE, Strollo PJ Jr (2014) Traditional and nontraditional cardiovascular risk factors in comorbid insomnia and sleep apnea. Sleep 37:593–600. doi: 10.5665/sleep.3506 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Cho YW, Kim KT, Moon HJ, Korostyshevskiy VR, Motamedi GK, Yang KI (2018) Comorbid insomnia with obstructive sleep apnea: clinical characteristics and risk factors. J Clin Sleep Med 14:409–417. doi: 10.5664/jcsm.6988 [DOI] [PMC free article] [PubMed] [Google Scholar]