Sleep Quality and Heart Rate Variability in Adolescents with Type 1 or Type 2 Diabetes

1. Introduction

Recent trends in the incidence rates of both type 1 (T1D) or type 2 diabetes (T2D) in the United States reveal increasing numbers in youth populations.¹ Along with this epidemiological evidence of younger individuals that will experience longer duration of diabetes, there is great concern for early identification of behavioral risks, such as poor sleep, for future complications. The American Diabetes Association’s (ADA) Standards of Medical Care (2021)² indicate that sleep patterns and duration should be included in routine clinical visits. However, there is limited evidence on the examination of sleep for adolescents with diabetes during these visits.

Over the past decade, dramatic effects of poor sleep quality on health outcomes for both youth and adults are noted. Insufficient sleep contributes to increased insulin resistance, hypertension, cardiovascular disease and mortality.³ In a large population-based study using data from adults included in the UK Biobank, having both a diagnosis of diabetes and sleep disturbances was associated with a greater risk of all-cause mortality than either condition alone.⁴ Of particular concern is the higher risk of mortality, mainly due to cardiovascular disease, reported in persons who are diagnosed with diabetes at a younger age compared to those diagnosed later in life.⁵

Currently, in comparison to youth, more is known about adults with T1D⁶ or with T2D⁷ that experience greater sleep disturbance than those without diabetes. Sleep disturbance for adults with diabetes consists of both impaired sleep quality and altered sleep duration.⁸ A recent meta-analysis found that poor sleep quality, as well as short and long sleep duration, is associated with less favorable glycemic control in adults with T2D.⁹ Further research reported significant associations between subjective poor sleep quality and lower heart rate variability (HRV) in adults with T2D.¹⁰ Heart rate variability is a well-validated measure of cardiovascular neurological modulation and decreases in HRV serve as the gold standard for early detection of cardiovascular risk and cardiovascular autonomic neuropathy (CAN),¹¹as well as an independent predictor of overall mortality.¹² The prevalence of CAN has been reported to be 12% in adolescents and young adults with T1D and 17% in those with T2D.¹³ Recent studies provide further evidence of the presence of cardiovascular risk factors in adolescents with T1D or T2D,^13,14 with risks more pronounced in those with T2D,^15,16 due to a more rapid progression of the pathophysiologic changes of insulin resistance and beta-cell failure than what occurs in adults with the disease.¹⁷

Although sleep studies in youth with T1D report poor sleep quality¹⁸ and greater glucose control with better self-reported sleep quality,¹⁹ less is known in this young cohort with T2D. To our knowledge, no studies have explored measures of sleep quality with heart rate variability, as an early indicator of cardiovascular disease risk in adolescents with diabetes. More evidence is needed to understand the possible association of sleep with known biomarkers for future cardiovascular disease, such as HRV, and the presence of differences that may exist for those with T1D versus T2D. Therefore, the primary hypothesis of this study was that poorer sleep quality would be associated with lower heart rate variability measures in adolescents with diabetes and that associations may differ by type of diabetes and glycemic control.

2. Material and Methods

2.1. Design, setting and sample

A secondary analysis of an existing parent dataset examining cardiovascular risks of adolescents with either T1D or T2D (PHS R01 NR07719) was conducted. In the original parent study, participants were recruited from a large, metropolitan pediatric diabetes clinic affiliated with a major academic medical center in the Midwestern United States. There were 101 adolescents with T1D and 37 with T2D included. Institutional review board approval was obtained. Parental consent and youth assent were received prior to enrollment in the study. Adolescents were eligible for inclusion in the study if they had a diagnosis of T1D or T2D for at least one year and were between 13 and 18 years of age. Adolescents were excluded if they had developed diabetes as a secondary condition to treatment for another chronic condition (e.g., cancer) or had a known cardiac defect. Data were collected at the clinical research center of the affiliated academic medical center.

2.2. Measures and methods

Sex and age-adjusted body mass index (BMI) and BMI z-scores were calculated based upon syntax files provided by the Centers for Disease Control (https://www.cdc.gov/growthcharts/percentile_data_files.htm). Glycemic control was determined by recent HbA_1c values established at the time of data collection and analyzed using the Abbott IMx® assay (Abbott Park, IL, USA).

The measure of sleep quality was derived from a single item taken from the Disease Impact subscale of the Diabetes Quality of Life Measure for Youth.²⁰ The Diabetes Quality of Life Measure for Youth is a well-validated instrument that was developed for the Diabetes Control and Complications Trial.²¹ Content validity of the Diabetes Quality of Life Measure for Youth was established by the independent review of a pediatric diabetologist, nurse practitioner, social worker, and educational psychologist, all of whom specialized in pediatric diabetes as well as by exploratory factor analysis and confirmatory structural modeling. The single item used in this secondary analysis included the question: How often do you have a bad night’s sleep? The possible responses on a Likert scale were as follows: 1=Never, 2=Very Seldom, 3=Sometimes, 4=Often, 5=All the Time. Other measures of sleep quality were not available in the existing dataset to be used in this secondary analysis.

Twenty four-hour heart rate variability (HRV) was measured by use of the 3-channel SpaceLabs Burdick Model 92510 digital Holter Recorder (Deerfield, WI, USA). Measurement and analysis of HRV are classified into frequency and time domain analyses. Kleiger et al.²² previously established the stability of both time and frequency domain HRV measures using 24-hour Holter monitoring, reporting correlation coefficients > .80.

The SpaceLabs Vision Premier^™ ECG Analysis and Editing software system utilizes the Fast Fourier Method of spectral analysis to calculate the frequency domain. Power spectral analysis of heart rate variability offers a highly sensitive source of data related to autonomic function by quantifying and discriminating between sympathetic and parasympathetic autonomic function through frequency of R-R variation. Abnormalities detected by power spectral analysis are more sensitive indicators of autonomic dysfunction, particularly in patients with diabetes who have diminished 24-hour R-R interval variability.²³

Use of Fast Fourier Transformations (FFT) provides a mathematical representation of the spectrum of frequency, called Hertz (Hz) or power. Total Hertz is the entire area under the curve in a power spectrum plot and represents the variance of all R-R intervals in the entire Holter recording. Frequency domain measures of total Hertz (0.01–1.00 Hz), low Hertz (0.04–0.15) and high Hertz (0.15–0.40) are converted to log transformation by the computer software to correct for skewness.²⁴ Low frequency (Hertz) activity (primarily sympathetic with some parasympathetic innervation) predominates during waking hours and high frequency (Hertz) activity (parasympathetic innervation) predominates during sleep.²⁵

Time domain analysis is computed on differing computations of the standard deviation of the beat-to-beat change in heart rate, based on normal sinus R-R intervals over time.²⁶ The time domain analysis of heart rate variability can be further divided into two categories. One category is derived from the R-R intervals, using means and standard deviations of the intervals measured in milliseconds. The standard deviation of all R-R intervals during a 24-hour period is the SDNN. In adults, SDNN is a standard measure of morbidity and mortality with values between 50 to 100 milliseconds reflective of compromised cardiac health and values that are less than 50 milliseconds associated with risks for sudden cardiac death.²⁷

The second category of time domain variables is derived from differences between adjacent R-R intervals and includes indices that are independent of circadian rhythms. Measures in this category are the pNN50 and the rMSSD. The proportion of the total R-R intervals that have differences of successive R-R intervals greater than 50 milliseconds represents pNN50. The square root of the mean squared differences of successive R-R intervals is the rMSSD. The pNN50 and rMSSD correlate highly with high-frequency power, reflecting parasympathetic modulation.²⁶

2.3. Statistical Analysis

Sample characteristics for demographic information, recent HbA_1c, blood pressure, body mass index (BMI), BMI z-score, Tanner stage of sexual maturation, duration of diabetes and sleep quality were computed using descriptive statistics of means, standard deviations and percentages. Linear regression analysis controlling for the effects of diabetes type, sex, body mass index using z-scores, sleep quality and HbA_1c on measures of HRV were computed. Since there is conflicting evidence to support that lower HRV is found in those with higher BMI²⁸ or adiposity,²⁹ BMI z-scores were included in the regression analysis. Associations between sleep quality and measures of HRV for the total sample, for those with T1D or T2D, and by sex within each group were also computed using Pearson correlation coefficients. Data were evaluated for normality and differences were examined using independent t-tests for comparisons of HRV by type of diabetes and by sex within diabetes type. The rationale to examine possible differences in HRV for females versus males within diabetes group was based upon earlier findings of lower HRV in female adolescents with³⁰ and without diabetes.³¹ The level of significance was set at 0.05. Data were analyzed using IBM SPSS Statistics for Windows Version 27.

3. Results

Demographic characteristics are presented in Table 1. There was a slightly higher representation of males versus females in the T1D group as compared with the T2D group. Racial variation between the groups was reflective of the known higher prevalence of T2D in minority groups. There were no differences in age, recent HbA_1c, or self-reported sleep quality between those with T1D versus T2D or between males versus females within each group. There were significant differences in diastolic blood pressure (BP), BMI, BMI z-scores, Tanner stage of sexual maturation and duration of diabetes between those with T1D versus T2D. Differences for BMI and BMI z-scores for males vs. females within groups (T1D or T2D) were non-significant.

Table 1.

Sample Characteristics

Variables	T1DM (n=101)	T2DM (n=37)	p value
Age (years), M ± SD	15.4 (± 1.9)	15.8 (± 1.9)	0.21
Sex
Male, n (%)	56 (55.4)	16 (43.2)
Female, n (%)	45 (44.6)	21 (56.8)
Race
Caucasian/White, n (%)	67 (66.3)	0 (0)
African American/Black, n (%)	29 (28.7)	26 (70.3)
Hispanic, n (%)	5 (5.0)	10 (27.0)
Asian, n (%)	0 (0)	1 (2.7)
Recent Hemoglobin A1c	8.7 ± 1.6	8.5 ± 2.5	0.61
Systolic Blood Pressure (mmHg)	113 ± 11	118 ± 12	0.06
Diastolic Blood Pressure (mmHg)	65 ± 9	69 ± 9	0.02
BMI, M ± SD	23.2 ± 3.8	34.3 ±9.3	< 0.001
Male	22.7 ± 4.1	35.1 ± 11.1	< 0.001
Female	23.9 ± 3.2	33.7 ± 7.8	< 0.001
BMI z-score, M ± SD	0.7 ± 0.8	2.0 ± 0.6	< 0.001
Male	0.6 ± 0.9	2.2 ± 0.6	<0.001
Female	0.9 ± 0.5	1.9 ± 0.6	<0.001
Tanner Stage, M ± SD	4.27 ± 0.9	4.67 ± 0.6	0.01
Duration of Disease (years), M ± SD	6.2 ± 3.7	2.2 ± 1.7	< 0.001
Sleep Quality, M ± SD	2.30 ± 1.0	2.40 ± 1.2	0.683
Male	2.18 ± .97	2.19 ± 1.0	NS
Female	2.44 ± .94	2.53 ± 1.3	NS

The results of the linear regression models are presented in Table 2. All models were significant at the p < 0.01 level of significance with adjusted R² values ranging from .095 to .199. Females and those with type 2 diabetes had significantly lower HRV measures, but there was no significant effect related to BMI z-scores or HbA_1c. Sleep quality was significantly associated with the HRV measures of high frequency, rMSSD and pNN50.

Table 2.

Regression Coefficients of Diabetes Type, Sex, BMI Z-Score, Sleep Quality on Heart Rate Variability

	B	SE	β	t	p	Adj. R2 **
AVE HR
DM type	4.579	2.321	.205	1.973	.05	.113
Sex	5.159	1.667	.261	3.096	.002
BMI z-score	−1.998	1.139	−.186	−1.754	.082
Sleep quality	1.184	.782	.124	1.513	.133
HbA1c	.527	.453	.098	1.165	.246
TOTFreq
DM type	−.450	.168	−.269	−2.683	.008	.169
Sex	−.431	.120	−.292	−3.591	.000
BMI z-score	.079	.062	.098	.960	.339
Sleep quality	−.100	.056	−.140	−1.762	.080
HbA1c	−.040	.033	−.100	−1.230	.221
HiFreq
DM type	−.622	.235	−.272	−2.649	.009	.132
Sex	−.372	.168	−.184	−2.218	.028
BMI z-score	.118	.115	.108	1.026	.307
Sleep quality	−.217	.079	−.223	−2.750	.007
HbA1c	−.034	.046	−.063	−.752	.454
LowFreq
DM type	−.454	.169	−.268	−2.689	.008	.177
Sex	−.423	.121	−.283	−3.498	.001
BMI z-score	.032	.083	.040	.392	.696
Sleep quality	−4.877	3.105	−.122	−1.571	.119
HbA1c	−.411	1.792	−.018	−.229	.819
SDNN
DM type	−22.398	9.223	−.239	−2.429	.017	.199
Sex	−29.489	6.597	−.357	−4.470	.000
BMI z-score	.921	4.520	.021	.204	.839
Sleep quality	−4.877	3.105	−.122	−1.571	.119
HbA1c	−.411	1.792	−.018	−.229	.819
rMSSD
DM type	−13.271	6.760	−.205	−1.957	.052	.095
Sex	−10.404	4.849	−.182	−2.146	.034
BMI z-score	4.975	3.323	.160	1.497	.137
Sleep quality	−6.570	2.283	−.238	−2.878	.005
HbA1c	1.036	1.317	.067	.786	.433
pNN50
DM type	−6.118	3.037	−.206	−2.015	.046	.139
Sex	−6.476	2.172	−.247	−2.981	.003
BMI z-score	2.806	1.488	.197	1.885	.062
Sleep quality	−3.153	1.023	−.249	−3.083	.002
HbA1c	−.009	.590	.001	−.016	.987

Note: DM type = diabetes mellitus type 1 (coded as 1) or type 2 (coded as 2); Sex coded as male (1) and female (2).

AVEHR = average heart rate over 24 hours;

TOTFreq = Total frequency variation of sympathetic and parasympathetic innervation in natural log (ms²);

HiFreq = Parasympathetic innervation in natural log (ms²);

LowFreq = Primary sympathetic with some parasympathetic innervation in natural log (ms²);

SDNN = Standard deviation in milliseconds of all R-R intervals:

rMSSD = Square root of the mean of the sum of squares of differences between adjacent R-R intervals;

PNN50 = Percentage of adjacent R-R intervals with greater than 50 ms difference

^**p < 0.01 for all models

Correlations between sleep quality and average heart rate and both time and frequency domain measures of HRV over 24-hours for the total sample, by diabetes type and by sex within group are shown in Table 3. These associations were stronger in those with T2D, particularly females. Differences in heart rate variability measures for those with T1D versus T2D, and for females compared to males within diabetes type are shown in Table 4.

Table 3.

Sleep Quality Correlations with Measures of Heart Rate Variability

	AVEHR	TOTFreq	HiFreq	LowFreq	SDNN	rMSSD	PNN50
Total Sample	0.169*	−0.190*	−0.256**	−0.185*	−0.175*	−0.260**	−0.282**
T1D
Total	0.072	−0.063	−0.127	−0.056	−0.068	−0.193*	−0.210*
Male	0.095	−0.105	−0.147	−0.123	−0.097	−0.161	−0.206
Female	−0.032	0.084	−0.063	0.124	0.094	−0.187	−0.161
T2D
Total	0.401*	−0.453**	−0.508**	−0.440**	−0.388*	−0.405*	−0.429**
Male	0.177	−0.218	−0.242	−0.264	−0.202	−0.189	−0.226
Female	0.499*	−0.587**	−0.654**	−0.521*	−0.547*	−0.519*	−0.637**

Note. Unadjusted correlations coefficients are shown.

AVEHR = average heart rate over 24 hours;

TOTFreq = Total frequency variation of sympathetic and parasympathetic innervation in natural log (ms²); HiFreq = Parasympathetic innervation in natural log (ms²);

LowFreq = Primary sympathetic with some parasympathetic innervation in natural log (ms²);

SDNN= Standard deviation in milliseconds of all R-R intervals;

rMSSD = Square root of the mean of the sum of squares of differences between adjacent R-R intervals;

PNN50 = Percentage of adjacent R-R intervals with greater than 50 ms difference

^*p < 0.05

^**p < 0.01

Table 4.

Heart Rate Variability Measures for Type of Diabetes and for Males versus Females by Diabetes Type

	AVEHR	TOTFreq	HiFreq	LowFreq	SDNN	rMSSD	PNN50
T1D	83 (10)	8.3 (0.7)**	6.7 (0.9)**	7.1 (0.7)**	154.4 (39.3)**	58.9 (28.3)	20.2 (12.9)
Male	81 (11)*	8.5 (0.7)a	6.8 (0.9)	7.2 (0.7)a	167.6 (37.9)a	63.4 (26.5)	22.6 (12.8)a
Female	85 (9) *	8.1 (0.6)b	6.5 (1.0)	6.9 (0.6)b	138.0 (34.9)b	53.4 (29.8)	17.1 (12.4)b
T2D	86 (9)	7.9 (0.7)**	6.2 (1.1)**	6.6 (0.8)**	129.4 (42.2)**	50.1 (28.6)	16.6 (13.6)
Male	80 (7)***	8.3 (0.7)a	6.6 (1.0)a	7.0 (0.6)a	146.2 (48.4)a	56.8 (29.3)	22.5 (15.8)a
Female	90 (8)***	7.6 (0.7)b	5.8 (1.0)b	6.3 (0.7)b	116.5 (32.3)b	45.1 (27.7)	12.2 (9.8)b

Note. Data are presented as mean scores and standard deviation in parentheses

^*p < 0.05 and

^**p < 0.01 for comparisons between types of diabetes

^***p < 0.001

Values within a column for sex comparisons within diabetes type with different superscripts (a and b) are significantly different at p < 0.05

AVEHR = Average heart rate over 24 hours;

TOTFreq = Total frequency variation of sympathetic and parasympathetic innervation in natural log (ms²);

HiFreq = Parasympathetic innervation in natural log (ms²);

LowFreq = Primary sympathetic with some parasympathetic innervation in natural log (ms²); SDNN = Standard deviation in milliseconds of all R-R intervals;

rMSSD = Square root of the mean of the sum of squares of differences between adjacent R-R intervals; PNN50 = Percentage of adjacent R-R intervals with greater than 50 ms difference

4. Discussion

To our knowledge, this is the first study to examine sleep quality and HRV in adolescents with T1D or T2D. In terms of the demographic findings, the slightly higher mean systolic and diastolic BP values in those with T2D compared to those with T1D were not clinically significant based upon the current recommendation for normal BP in youth ≥ 13 years of age to be < 120 mmHg systolic and < 80 mmHg diastolic.³² The higher mean values for BMI, BMI z-scores, Tanner stage and duration were expected due to the known associations of overweight or obesity and later onset age in those with T2D.

Our findings support that poor quality of sleep is associated with lower measures of HRV in adolescents with diabetes, specifically measures that represent parasympathetic modulation that dominates during normal sleep. There is the need for an increased focus on routine assessment of sleep and early signs of cardiovascular autonomic changes, such as increased heart rate at rest during examination, and interventions for improvement in preventative strategies, such as better sleep hygiene in adolescents with diabetes. The lower HRV measures of females compared to males within each diabetes group is consistent with other research on HRV in adolescents with diabetes.³⁰ Although the possible effect of higher BMI on the relationship of sleep with HRV was not established with our findings, this warrants further examination in larger samples of adolescents with T1D or T2D since overweight is a concern now for those with either type of diabetes.

Although better glycemic control has been associated with sleep quality in previous research,¹⁹ we did not find any association in our study based upon self-report of sleep quality with HbA_1c. Research reported by Perfect et al.³³ also did not find associations between self-reported sleep variables and glycemic control. Recent investigations of youth with T1D using continuous glucose monitoring (CGM) and actigraphy or polysomnography were completed with small samples (n ≤ 40) and did find significant associations between glucose control and sleep parameters.^33–35 Higher CGM values were associated with greater time in stage N2 (light) sleep^33,35 and greater glucose variability was also noted with increased sleep pattern variability.³⁴ Therefore, the use of CGM and more objective measures for sleep may offer greater sensitivity for detecting associations among these variables in future research.

In youth with T1D followed in the Search for Diabetes in Youth Study, a decline in HRV was associated with worsening glycemic control over a 6-year period independent of demographic and traditional cardiovascular disease risks.³⁶ Recently, cohort comparisons of youth with T1D or T2D from 2002 through 2019 revealed that current youth and young adults with T1D in the United States have worse glycemic control than earlier cohorts, despite the increased availability of newer therapies, more aggressive glycemic targets over time and greater availability of diabetes technologies for self-management.³⁷ The same study found that individuals with T2D diagnosed in childhood and with a duration of disease over 10 years also exhibited worsening glycemic control in recent years. Although the findings of our study did not support any effect of glucose control on HRV for adolescents with T1D or T2D, further study is needed on the contribution of poor glycemic control and poor quality sleep on future complications.

The finding of stronger associations of difficulty with sleep and lower HRV in females with T2D compared to other groups raises an alert for early identification of risks in this young cohort. Current research evidence indicates that cardiovascular disease, including coronary heart disease and stroke, affects women with diabetes more than men.^38,39 Women with diabetes generally have more cardiovascular risk factors, including a higher prevalence of obesity, dyslipidemia, hypertension, and waist to hip ratios.⁴⁰ The risk of heart failure and diabetes-associated cardiomyopathy is reported to be greater in women versus men with diabetes.⁴¹ A recent meta-analysis assessing mortality due to cardiovascular disease between men and women with T2D found that women had a higher all-cause mortality rate (HR 1.17; 95% CI 1.12–1.23), as well as higher mortality rates specifically related to cardiovascular disease (HR 1.97, 95% CI: 1.49–2.61).⁴²

Cardiovascular disease in men and women with pre-diabetes has also been compared. Ahn et al.⁴³ found that women started to have an increased risk for cardiovascular disease when diagnosed with pre-diabetes, however men did not start to develop this risk until their pre-diabetes progressed to diabetes. This may suggest that cardiovascular disease occurs earlier in the diabetic disease trajectory for women when compared to men.

As a marker for cardiovascular disease risk, lower HRV as a measure of reduced cardiac autonomic function is found to differ for women versus men with diabetes.^44,45 Nolan et al.⁴⁵ found that lower heart rate variability was associated with the duration of diabetes in men only, while reduced heart rate variability was associated with the age of diagnosis in women.

The variation between women and men in the relationship between glucose variability and autonomic dysfunction has also been studied in those newly diagnosed with diabetes. Fleisher et al.⁴⁵ found that glucose variability, the wide fluctuation in glucose levels, was associated with cardiac autonomic dysfunction in women only, not men, even when factors including glycemic control, cholesterol levels, and smoking were taken into account.

Lack of control or management of cardiovascular risk factors has been hypothesized to contribute to the increased risk of cardiovascular disease in women.³⁸ In one study, screening for cardiovascular disease appeared to occur 12% less often in women when compared to men.⁴⁶ Blood pressure and cholesterol, common risk factors associated with cardiovascular disease, are noted to be less well-controlled in women versus men with diabetes.^40,47 Women may also have higher rates of depression than men, which can result in non-adherence and result in poor control of risk factors.³⁹ Together this reduction in screening, treatment and management of other cardiovascular risk factors, and non-adherence may all contribute to women being more adversely affected by cardiovascular disease when compared to their male counterparts with diabetes. The research evidence available in adult studies highlighting the risks for heart disease for women with diabetes serves as warning signs that more research attention is needed in much younger cohorts so that interventions can be targeted to strategies for minimizing these risks and future complications. With the earlier onset of T2D in younger cohorts and the awareness of potential sex variation in cardiovascular risks with a diagnosis of diabetes, more studies are needed to explore differences between females and males, responses to treatment and longitudinal follow-up.

5. Limitations

The generalizability of this study is limited since the results are based on a secondary analysis of an existing dataset originating from a convenience sample of a single clinic site and there were unequal numbers of adolescents with T1D or T2D. Findings of this study are not predictive and provide a glimpse into the possible influence of sleep quality on HRV measures and differences for adolescent females versus males with diabetes. The use of self-report data for sleep quality is also a less rigorous measurement and more robust assessments of both subjective and objective measurement of sleep are needed to fully evaluate sleep behaviors in this young cohort. More comprehensive and valid measures of sleep quality are available for use with adolescents, such as the Pittsburgh Sleep Quality Index (PSQI)⁴⁸ and the Adolescent Sleep–Wake Scale (ASWS),⁴⁹ The study results provide preliminary evidence of the importance of conducting this research in a larger sample for further examination of the influence of sleep on HRV in adolescents with T1D or T1D and the possible variations for females versus males.

6. Conclusions

There is a tremendous urgency to raise awareness for optimal sleep hygiene that may minimize future disease risks for adolescents with diabetes. Female adolescents with diabetes, particularly those with T2D, may have greater risks related to poor sleep quality. Poor quality sleep and lower HRV are particular concerns that warrant further investigation along with more objective measures of both sleep quality and sleep duration to better understand key vulnerabilities in youth populations with diabetes.