Sleep characteristics in type 1 diabetes and associations with glycemic control: systematic review and meta-analysis

Abstract

Objectives: The association between inadequate sleep and type 2 diabetes has garnered much attention, but little is known about sleep and type 1 diabetes (T1D). Our objectives were to conduct a systematic review and meta-analysis comparing sleep in persons with and without T1D, and to explore relationships between sleep and glycemic control in T1D.

Methods: Studies were identified from Medline and Scopus. Studies reporting measures of sleep in T1D patients and controls, and/or associations between sleep and glycemic control, were selected.

Results: A total of 22 studies were eligible for the meta-analysis. Children with T1D had shorter sleep duration (mean difference [MD] = -26.4 minutes; 95% confidence interval [CI] = -35.4, -17.7) than controls. Adults with T1D reported poorer sleep quality (MD in standardized sleep quality score = 0.51; 95% CI = 0.33, 0.70), with higher scores reflecting worse sleep quality) than controls, but there was no difference in self-reported sleep duration. Adults with TID who reported sleeping >6 hours had lower hemoglobin A1c (HbA1c) levels than those sleeping ≤6 hours (MD = -0.24%; 95% CI = -0.47, -0.02), and participants reporting good sleep quality had lower HbA1c than those with poor sleep quality (MD = -0.19%; 95% CI = -0.30, -0.08). The estimated prevalence of obstructive sleep apnea (OSA) in adults with TID was 51.9% (95% CI = 31.2, 72.6). Patients with moderate-to-severe OSA had a trend toward higher HbA1c (MD = 0.39%, 95% CI = -0.08, 0.87).

Conclusion: T1D was associated with poorer sleep and high prevalence of OSA. Poor sleep quality, shorter sleep duration, and OSA were associated with suboptimal glycemic control in T1D patients.

Keywords: Glycemic control; Meta-analysis; Obstructive sleep apnea; Sleep duration; Sleep quality; Type 1 diabetes.

Fig. A1.

Relationship between sleep stages and glycemic control in type 1 diabetes (T1D) patients. (A) Mean difference of percentages of sleep time spent in light sleep between participants with optimal (HbA1c <7%) and suboptimal (HbA1c ≥7%) glycemic control (GC) (calculated by percentage of sleep time of participants with optimal GC minus those with suboptimal GC). (B) Mean difference of percentages of sleep time spent in deep sleep between those with optimal and suboptimal GCs. NREM, non-rapid eye movement; REM, rapid eye movement.

Fig. A2.

Relationship between sleep duration and glycemic control (GC) in adults with type 1 diabetes (T1D). (A) Mean difference in HbA1c levels between participants with longer sleep duration (>6 hours) and those with shorter sleep duration (≤6 hours). (B) Mean difference in sleep duration between participants with optimal (HbA1c <7%) and suboptimal (HbA1c ≥7%) GCs (calculated by sleep duration in minutes of those with optimal GC minus those with suboptimal GC). PSG, polysomnography.

Fig. A3.

Relationship between sleep duration and glycemic control (GC) in children with type 1 diabetes (T1D). (A) Mean difference in HbA1c levels between participants with longer and shorter sleep durations, calculated by HbA1c in those with longer sleep duration minus that of those with shorter sleep duration. (B) Mean difference in sleep duration between participants with optimal (HbA1c < 7.5–8%) and suboptimal (HbA1c ≥ 7.5–8%) GCs.

Fig. A4.

Relationship between sleep quality and glycemic control (GC) in adults with type 1 diabetes (T1D). (A) Mean difference in HbA1c levels between participants with good sleep quality (sleep efficiency ≥85% as measured by polysomnography [PSG] or actigraphy, or per sleep quality score cutoff according to the sleep questionnaire used) and those with poor sleep quality. (B) Mean difference in sleep efficiency between participants with optimal (HbA1c <7%) and suboptimal (HbA1c ≥7%) GCs.

Fig. A5.

Relationship between obstructive sleep apnea (OSA) and glycemic control (GC) in patients with type 1 diabetes (TID). (A) Mean difference in HbA1c levels between participants with OSA and without OSA in adults (calculated by HbA1c in those with OSA minus those without OSA). (B) Mean difference in HbA1c levels between those with moderate to severe OSA (AHI ≥15) and those without OSA (AHI <5) in adults (calculated by HbA1c in those with moderate to severe OSA minus those without OSA). (C) Mean difference in AHI between those with optimal (HbA1c < 7%) and suboptimal (HbA1c ≥ 7%) GCs (calculated by AHI of those with optimal GC minus those with suboptimal GC).

Fig. A6.

Funnel plots of the mean difference between patients with type 1 diabetes (T1D) and control participants. (A) Sleep duration. (B) Sleep quality.

Fig. A7.

Funnel and contour-enhanced funnel plots for mean differences between adult type 1 diabetes (T1D) patients with good and poor glycemic controls. (A) Sleep duration as obtained by objective measurements. (B) Sleep quality by objective measurements.

Fig. 1.

Flow chart of study selection. Poolings were performed when there were three or more studies in the same category.

Fig. 2.

Mean difference in sleep duration between patients with type 1 diabetes (T1D) and controls (calculated by sleep duration in minutes of T1D patients minus that of controls). (A) Adults by questionnaire. (B) Adolescents/children by polysomnography.

Fig. 3.

Comparisons of sleep quality between patients with type 1 diabetes (T1D) and controls. (A) Mean difference in sleep efficiency by polysomnography (PSG) (sleep efficiency of T1D patients minus that of controls). (B) Standardized mean difference in sleep quality score by questionnaire with higher score reflecting worse quality (T1D patient score minus that of controls). (C) Association between T1D and good sleep quality.