SEARCH

SEARCH BY CITATION

Keywords:

  • diabetes;
  • exenatide;
  • sleepiness

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure Statement
  9. References

We investigate the effects of exenatide on excessive daytime sleepiness (EDS), driving performance and depression score in patients with type 2 diabetes with EDS. Eight obese patients with diabetes but without obstructive sleep apnoea (OSA) participated in a placebo-controlled single-blind study during which multiple wakefulness and sleep latency test, Epworth score, driving performance, depression score, fasting glucose and glycated haemoglobin (HbA1c) levels were assessed at baseline, end of placebo and treatment phase at baseline and after 22 weeks of treatment. Mean (±standard error of the mean) age, body mass index (kg m2) and HbA1c [mmol mol−1 (%)] of patients at baseline were 50 ± 4.9 years, 37.6 ± 1.1 and 65 ± 19 (8.06 ± 0.41), respectively. When compared to placebo, exenatide treatment was associated with a decrease in both subjective and objective sleepiness, based on the Epworth score reduction and the sleep latency increase assessed by multiple objective sleepiness and sustained attention (OSLER) tests, respectively. Mean sleep latency time (adjusted for change in HbA1c and weight) were 32.1 ± 1.7, 29.1 ± 1.7 and 37.7 ± 1.7, respectively (= 0.002). Modelling for covariates suggested that improvement in mean sleep latency time is predicted by changes in weight (= 0.003), but not by changes in HbA1c (= 0.054). Epworth sleepiness score was reduced significantly (values for placebo versus exenatide: 11.3 ± 1.2 versus 5.7 ± 1.3; = 0.003). No significant change was noted in the depression score and driving performance. Exenatide is associated with a significant reduction in objective sleepiness in obese patients with type 2 diabetes without OSA, independent of HbA1c levels. These findings could form a basis for further studies to investigate the pathophysiological mechanisms of sleepiness in obese patients with type 2 diabetes.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure Statement
  9. References

Excessive daytime sleepiness (EDS) and fatigue are common complaints among patients with diabetes and constitute an important health and social issue. EDS is also a cardinal feature of obstructive sleep apnoea (OSA), and in view of the high prevalence of OSA in patients with diabetes (West et al., 2006) and its independent association with the metabolic syndrome (Gruber et al., 2006; Newman et al., 2001), complaints of EDS among patients with diabetes are often assumed to be due to the presence of underlying OSA. Various studies, however, have shown that obesity (Resta et al., 2003; Vgontzas et al., 1998), depression (Breslau et al., 1997; Ford and Kamerow, 1989) and diabetes (Resnick et al., 2003) are responsible for EDS, independent of OSA. In a large random population sample with selected patients evaluated in the sleep laboratory, depression was shown to be the most significant risk factor for EDS, followed by high body mass index (BMI), diabetes and, finally, OSA (Bixler et al., 2005). In our study of patients with diabetes not known to have OSA, increased BMI was shown to be an independent predictor of EDS while glycated haemoglobin (HbA1c) was significantly higher in patients with EDS compared with patients without EDS (Iqbal et al., 2007). Thus, the mechanism of EDS in diabetes appears to be multi-factorial and is associated with obesity and other factors independent of OSA.

Evidence from studies on EDS or in experimentally induced EDS after sleep deprivation has suggested that the independent association between obesity and EDS is mediated by increased production of proinflammatory ‘somnogenic’ adipose tissue-derived cytokines (Späth-Schwalbe et al., 1998; Vgontzas et al., 2004). The significance of EDS, daytime somnolence and fitness to drive within the context of poor glycaemic control has not been explored. Increased understanding of the effects of concurrent glucose and weight reduction on EDS, driving performance and depression scores among obese patients with diabetes with EDS may therefore have important therapeutic and social implications.

The glucagon-like peptide (GLP)-1 receptor analogue exenatide is a recognized treatment for hyperglycaemia in obese patients with type 2 diabetes (NICE, 2009). Clinical studies have shown efficacy in glucose-lowering and consistent reductions in weight (Buse et al., 2007). We aim to investigate whether the administration of exenatide to obese patients with type 2 diabetes without OSA would result in a significant improvement in objective measure of daytime sleepiness. We also aim to evaluate the effect of exenatide on levels of somnogenic adipocytokines, glycaemic control, weight and driving performance.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure Statement
  9. References

All studies were performed at the Sherwood Forest Hospitals Foundation Trust. Eight male obese (BMI 28–40) subjects with type 2 diabetes, on a stable medication regimen (>6 months) and with HbA1c >53 mmol mol−1 (7.0%) were recruited. All subjects underwent a sleep study in the sleep laboratory to exclude OSA. Patients with congestive heart failure, depression, cancer, central nervous system demyelinating disorders, recurrent or active infection, very poorly controlled diabetes [HbA1c >108 mmol mol−1 (12%)] or receiving anti-obesity agents were excluded from the study. The study received ethical approval from the Nottingham Research Ethics committee and clinical trial authorization from the Medicines and Healthcare products Regulatory Agency (MHRA) (ref. 09/H0408/70; MHRA licensing authority reference number: 27699/0001/001-0001). All participants provided written consent to the study.

This was an exploratory placebo-controlled study lasting for 22 weeks, during which eligible participants were assessed immediately before the administration of placebo (baseline), 1 week after a 10-week administration of saline (placebo) and 1 week after a 10-week administration of (exenatide) treatment. Patient and respiratory physiologists, but not the study investigators, were blinded to treatment regimen. Exenatide therapy was initiated at a 5 μg twice-daily dose by subcutaneous injection and increased to 10 μg twice daily within 4 weeks of treatment initiation. Assessments included testing for objective sleepiness and sustained attention (OSLER test) (Krieger et al., 2004), driving performance (driving simulator), biochemical markers [HbA1c, lipids, tumour necrosis factor (TNF)-α and interleukin (IL)-6 levels], clinical parameters (BMI, waist, systolic and diastolic blood pressure), completion of a food and exercise diary, completion of a questionnaire to assess daytime sleepiness (Epworth Sleepiness Scale), assessment of depressive mood (Beck Depression Inventory scale) and a quality of life questionnaire (SF36). We used a fixed-order design versus a traditional cross-over design for three reasons: (i) to control carry-over effects of the drugs on sleep and sleepiness measures, which in a cross-over design would have required a several-week washout period and/or a significantly larger sample; (ii) to minimize the possibility of change in baseline symptomatology due to a prolonged washout period; and (iii) to expose the minimum of subjects to an experimental condition with untried outcomes. One limitation of the fixed-order design is that it does not control for order effects. All respiratory and sleep assessments were performed by trained cardiorespiratory technician.

OLSER test procedure

The OSLER test consisted of a 40-min sleep-resistance challenge conducted in a dark room isolated from external noise. The subject, lying in a semi-recumbent position, was asked to remain awake without using specific strategies. By hitting a button placed on a box directly connected to a personal computer, the subject was instructed to respond to a visual stimulus (light-emitting diode flash) which appeared for 1 s every 3 s. A total of 800 stimulations per test were emitted. All tests were video-recorded to verify that the subjects were following instructions. Each session ended automatically after 40 min or before if the subject did not give a response to seven consecutive flashes (i.e. 21 s, representing approximately the minimal time needed to score a conventional epoch of sleep). A sleep latency (test duration) was determined for each session. The percentage of time corresponding to the appearance of errors during the OSLER test was analysed [(3 s × number of omissions/sleep latency duration in seconds) × 100]. Each subject underwent the OSLER test three times (at 09:00, 11:00, 13:00 and 15:00 hours) allowing the assessment of vigilance at different times of the day. The first OSLER session (09:00 hours) was started on an average of 2 h after the subjects had awakened.

Driving simulator

To assess driving objectively in a controlled environment, the Divided Attention Steering Simulator (DASS) (Stowood Scientific Instruments, Oxford, UK) was used. DASS performs two types of divided attention driving test—the two-dimensional ‘turbulence’ test (George, 1996) and the three-dimensional ‘road’ test (Land and Horwood, 1995). Both tests are validated in patients with sleep apnoea compared to normal controls. The two-dimensional test simulated steering down a motorway while being buffeted by side winds. The patient is required to steer the car while a forcing function is applied to move the car pseudo-randomly from side to side. Centre analysis result is reported as the mean error from centre (mean deviation of the centre of the car from the centre of the road). The three-dimensional road test is designed to emulate steering along a curved road. Off-road events of >15 s terminate the test. Curve analysis is reported as mean error curve (mean difference between the steering angle and the road angle). For both analyses, the average response time is recorded as the average time (in seconds) it took the user to respond to the target number. All patients had exposure to DASS prior to commencement of the simulator to ensure that they were familiar with the system.

Sleep study assessment

All patients underwent a standard sleep study to exclude sleep apnoea as per routine clinical practice, using the minimal patient contact sleep diagnosis system [VISI-3; Stowood Scientific Instruments Ltd (SSI),], which included the following parameters: digital video/audio, calibrated sound level measurement, electrocardiography, R–R timing, pulse transit time, airflow, respiratory effort, body position and movement. Oxyhaemoglobin saturation (SaO2) was measured via pulse oxymetry. A software package was used for downloading, viewing, reporting and analysis of trend recording oximeters and capnometers. Continuous nasal airflow delivery was by the Horizon nasal continuous positive airway pressure (CPAP) system. OSA was excluded if either of the following criteria using the apnoea–hypopnoea index (AHI) or respiratory disturbance index (RDI) were not met: (i) AHI or RDI greater than or equal to 15 events per hour, or (ii) AHI or RDI ≥ 5 and ≤14 events per hour with documented symptoms of excessive daytime sleepiness, impaired cognition, mood disorders or insomnia, or documented hypertension, ischaemic heart disease or history of stroke.

Statistical analysis

For data management and statistical analysis, spss for Windows was used. The mixed-effects model was applied to assess pairwise differences between baseline, placebo and treatment values for outcomes of interest. We used regression analysis to compare the means for baseline, placebo and treatment and adjusted for values at baseline and covariates (e.g. changes in HbA1c, weight and depression scores, whichever was appropriate depending on outcome variables of interest). The chi-square test was used for qualitative variables. Where comparisons between baseline and placebo did not show any significant differences in outcome variables, we presented data for treatment versus placebo comparisons. Based on the study by Vgontzas et al., to detect differences in sleep latency of 3.1 with standard error of 1.0 in a paired analysis we would require eight patients for a 76% study power with an α = 0.05. Results are expressed as mean (±standard error of the mean).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure Statement
  9. References

The mean age of patients involved in the study was 50 ± 4.9 years. Mean BMI (kg m2) and HbA1c (%) at baseline were 37.6 ± 1.1 and 65 ± 19 mmol mol−1 (8.06 ± 0.41%), respectively. Two of eight patients experienced transient nausea as an adverse effect during the active treatment phase, resolving after 4 weeks of treatment. There was a trend towards reduction in HbA1c, weight and BMI, but all parameters did not reach statistical significant. Epworth score (subjective assessment of EDS) was reduced significantly following 10 weeks’ treatment with exenatide (pairwise comparison showed = 0.001 for baseline versus post-treatment Epworth and = 0.001 between post-placebo versus post-treatment Epworth. Patients’ clinical and metabolic parameters are summarized in Table 1.

Table 1. Patient’s clinical parameters across three conditions (baseline, end of placebo phase and end of treatment phase)
 BaselinePlaceboExenatide P-value
  1. BMI, body mass index; HbA1c, glycated haemoglobin; LDL, low-density lipoprotein.

  2. Clinical parameters across three conditions (baseline, end of placebo phase and end of treatment phase) for patients involved in the study (= 8). Results are expressed as mean (±standard error of the mean). The mean age of patients involved in the study was 50 ± 4.9 years. P-value denotes analysis between placebo and exenatide and between baseline and placebo (in brackets).

HbA1c (IFCC mmol mol−1) %64 (19) 8.01 (0.4) 70 (19) 8.56 (0.4) 62 (0.19) 7.84 (0.4) 0.46 (0.30)
Weight (kg)113.8 (5.9)113.7 (5.8)94.8 (6.2)0.38 (0.98)
BMI (kg m2)37.6 (1.1)37.5 (1.0)31.1 (5.3)0.280 (0.97)
Waist circumference (cm)122.0 (4.4)121.8 (4.5)116.2 (5.6)0.48 (0.97)
Total cholesterol (mmol l−1)4.6 (0.3)4.7 (0.3)4.4 (0.2)0.44 (0.79)
LDL cholesterol (mmol l−1)2.5 (0.2)2.6 (0.2)2.4 (0.2)0.34 (0.70)
Triglyceride (mmol l−1)2.2 (0.3)2.7 (0.7)1.9 (0.1)0.32 (0.52)
Transaminase enzyme (mmol l−1)33.9 (8.7)41.2 (11.7)40.7 (7.6)0.97 (0.64)
Epworth score12.3 (1.2)11.3 (1.2)5.7 (1.3)0.003 (0.58)

Daytime sleepiness (OSLER test)

A significant increase in sleep latency was observed between treatment with placebo and exenatide, and the difference was significant after adjusting for changes in HbA1c, weight and depression scores. Mean sleep latency times (adjusted for change in HbA1c and weight) were 32.1 ± 1.7, 29.1 ± 1.7 and 37.7 ± 1.7 min for baseline, placebo and exenatide phase of the study (= 0.002). Pairwise comparison showed no difference between baseline and placebo values (= 0.207), but a significant increase between baseline and exenatide (= 0.021) and between placebo and exenatide values (= 0.001). Modelling for covariates suggested that improvement in mean sleep latency time is predicted by changes in weight (= 0.003), but not influenced by changes in HbA1c (= 0.054).

For the OSLER analysis at baseline and following treatment with placebo, sleep latency decreases over the day as determined by the four time-points of the OSLER analysis performed during the day (09:00, 11:00, 13:00 and 15:00 hours). Sleep latency was lowest at the final time-point (15:00 hours) for baseline and after placebo: mean sleep latency times were 23.8 ± 3.5 and 21.5 ± 3.5 min, respectively. Following treatment with exenatide sleep latency increases throughout the day, higher at each time-point, and maximum at OSLER analysis at 15:00 hours (40.0 ± 3.6 min) (Fig. 1). Overall, there was a significant improvement in sleep latency following adjustment for HbA1c and weight change between each time-point (= 0.028), with pairwise comparison identifying significant increase in sleep latency at all time-points after 09:00 hours between placebo and exenatide treatment (Fig. 2). Again, weight reduction but not HbA1c reduction contributed to the increase in sleep latency across the time-points between baseline and exenatide and between placebo and exenatide.

image

Figure 1.  Mean sleep latency (min) across all four time-points during daytime study of the objective sleepiness and sustained attention (OSLER) test. Each bar represents the mean ± standard error; *< 0.05, adjusted change between placebo and exenatide.

Download figure to PowerPoint

image

Figure 2.  Sleep latencies (min) for all four individual time-points during daytime study of the objective sleepiness and sustained attention (OSLER) test. Because adjusted mean change between baseline and placebo were not statistically significant, only sleep latency tests for placebo and exenatide treatment, adjusted for covariates, are shown. Data represent mean sleep latency time. All point differences between placebo and exenatide at 11:00, 13:00 and 15:00 hours were statistically significant (< 0.05).

Download figure to PowerPoint

Depression (Beck Depression Inventory) and quality of health (SF36 health status questionnaire) score

There was a significant reduction in depression scores from baseline to exenatide groups, but a non-significant reduction in depression scores between placebo and exenatide (Fig. 3), which persisted after adjustment for HbA1c and weight change (= 0.001). Regression analysis showed that both HbA1c (= 0.063) and weight change (0.155) did not contribute to the reduction in patients’ depression score observed during the study. No significant change was noted across all scores of the SF36 health status questionnaire.

image

Figure 3.  Change in Beck Depression Inventory score between baseline, placebo and exenatide treatment. Data represent mean ± standard error, < 0.05 between baseline and exenatide values.

Download figure to PowerPoint

Driving performance

No significant difference was noted in the average drivers’ reaction times (s) between placebo and exenatide treatment (2.46 ± 0.08 versus 2.36 ± 0.17, = 0.62). Similarly, analysis of mean error curve and mean error for centre tracking was not significantly different between placebo and exenatide treatment (0.32 ± 0.03 versus 0.33 ± 0.04, = 0.76) and (0.29 ± 0.06 versus 0.29 ± 0.06, = 0.96).

IL-6, TNF-α and C-reactive protein

There were no differences in both cytokine levels during the course of the study. Compared with placebo, levels of IL-6 (pg ml−1) and TNF-α (pg ml−1) after 10 weeks of exenatide treatment were (2.9 ± 0.5 versus 3.8 ± 0.9; = 0.47) and (3.2 ± 0.6 versus 4.4 ± 0.7; = 0.29). Similarly, levels of C-reactive protein (mg l−1) were not significantly different between placebo and exenatide groups (7.1 ± 2.0 versus 7.7 ± 2.0; = 0.87).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure Statement
  9. References

This study is the first to demonstrate that pharmacological intervention which aims to target simultaneous reductions in weight and improvement on glucose control is associated independently with a significant reduction in objective and subjective sleepiness/wakefulness in obese patients with type 2 diabetes without OSA. The reduction in sleepiness appears to be robust, and was significant after adjustment for placebo values as well as reductions in HbA1c. The favourable effect of exenatide on sleepiness, however, did not influence driving performance.

Large-scale observational studies have shown a strong association between BMI, diabetes and depression with EDS in the absence of OSA (Bixler et al., 2005; Resta et al., 2003). The aetiology of EDS among obese patients with diabetes is therefore likely to be complex and multi-factorial. Using regression analysis, we found that reduction in weight but not improvement of HbA1c levels were significant predictors of reduced EDS. This finding is indirectly supportive of a previous hypothesis which implicated proinflammatory cytokines derived from adipose tissue stores, IL-6 and TNF-α, to be mediators of excess daytime sleepiness and fatigue in human (Vgontzas et al., 1999, 2004). Our study, however, showed that the objective (OSLER) and subjective (Epworth score) reduction in EDS induced by exenatide was not associated with significant reductions in the levels of IL-6 or TNF-α, although we accept that multiple measurements of these cytokines are likely to be required to confirm this negative association.

An alternative mechanism of EDS is explained by recent understanding on the neural pathways which control sleep and arousal. A major sleep-promoting nucleus is the gamma-aminobutyric acid (GABA)-ergic ventrolateral–pre-optic nucleus in the hypothalamus (Szabadi, 2006). Ischaemia in the perifornical area in the hypothalamus, for example, has been implicated in the mechanism of sleepiness (Nishino, 2011) while loss of orexin/hypocretin signalling is associated with chronic sleepiness (Mochizuki et al., 2011). As GLP-1 analogue is also a brain neuropeptide with diverse central action and receptor distribution in the hypothalamus (Baggio et al., 2004; Parkinson et al., 2009), it is tempting to speculate that improvement in EDS with exenatide could have been mediated centrally via a direct action of exenatide on the hypothalamus. This hypothesis may also explain the trend towards reduction in depression scores seen in this study, as well as the previously reported improvement in wellbeing and quality of life among patients prescribed exenatide (Best et al., 2011). Further studies are needed in order to clarify the central effects of exenatide in regulating sleepiness and wellbeing via neuronal pathways in the hypothalamus. Importantly, we observed that driving performance was not affected by EDS at baseline and did not change following treatment with exenatide.

Several limitations to this study need to be highlighted. Unfortunately, usual duration and quality of sleep at home is not known. Specifically, no sleep data are available at baseline and during the different phases of the study. Similarly, information on sleep duration or quality before the baseline and final assessments were not available and may limit data interpretation. Despite a trend towards weight reduction between placebo and exenatide treatment, this did not reach statistical significance. The observed weight reduction in this study was much higher than that seen in clinical trials of exenatide. Results from the UK nationwide Association of British Clinical Diabetologists (ABCD) (Ryder et al., 2010) audit, however, suggest that the use of exenatide was associated with a wide variation in weight loss between different subjects, with many individuals experiencing weight loss of >20 kg. We therefore believe that the patients recruited to this study represented a highly motivated group of subjects who supplemented lifestyle changes with exenatide therapy to induce weight loss. No specific advice, however, was provided by the investigators towards lifestyle or dietary changes in this study. Finally, the sample size was small, which limits the full generalization and covariate analysis of this study. Nevertheless, we believe that the findings of this study are novel and may form a basis for a larger, more definitive study in the future.

In conclusion, this is the first interventional study to investigate the role of exenatide in modulating daytime sleepiness and driving performance in obese patients with type 2 diabetes without OSA. We showed that exenatide improved EDS significantly independent of HbA1c reduction. The mechanisms for these observations are unknown, and may be explained only partially by weight reduction. We hypothesized, however, that the accumulating evidence on the role of hypothalamic nuclei in regulating sleep and wakefulness, as well as the well-recognized pleotropic effects of exenatide in conjunction with the presence of GLP-1 receptors in the hypothalamus, would suggest a direct effect of exenatide in affecting sleepiness. Larger interventional studies and further understanding of the role of exenatide in affecting the neuronal pathway involved in sleepiness will promote our understanding of the pathophysiological mechanisms of sleepiness in obese patients with type 2 diabetes. This may lead to more effective treatment of this highly prevalent problem among this challenging group of patients.

Acknowledgement

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure Statement
  9. References

This study was funded by Eli Lilly Pharmaceuticals.

Disclosure Statement

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure Statement
  9. References

The study was funded by Eli Lilly pharmaceuticals. None of the investigators received any personal fees towards their work in this study. Eli Lilly had no access to the data generated, the analysis or the interpretation of the data from this study.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure Statement
  9. References