Effect of continuous positive airway pressure therapy on glucose control

Abstract

Obstructive sleep apnea (OSA) and diabetes mellitus are both highly prevalent disorders. There has been a recent recognition of an association between insulin resistance and sleep apnea. Continuous positive airway pressure (CPAP) has emerged as an effective therapy for treatment of OSA and has been shown to positively influence numerous pathophysiological factors that contribute to cardiovascular risk. There is emerging data that explores the influence of CPAP therapy, insulin sensitivity and glycemic control. In the current review, we examine this literature critically and formulate a synopsis that summarizes the current knowledge in this field.

INTRODUCTION

Obstructive sleep apnea (OSA) is characterized by episodic and repetitive upper airway narrowing during sleep that leads to a well recognized clinical syndrome of snoring and excessive daytime sleepiness[1]. OSA has been independently associated with hypertension and cardiovascular disease[2]. A high prevalence rate is being increasingly recognized, particularly in patients with metabolic syndrome, and several studies have suggested an independent association between OSA and insulin resistance (IR) and glycemic control[3-11].

EPIDEMIOLOGY

A recent study from France followed 806 elderly subjects over 7 years and found an independent association between sleep apnea and metabolic syndrome, even after correcting for age, gender and obesity[11]. A prospective study from Japan that spanned five years, examined nocturnal oximetry in over 4000 patients and calculated that a multivariable-adjusted hazard ratio (95% CI) for developing type 2 diabetes was 1.69 (1.04-2.76) among those with moderate to severe nocturnal intermittent hypoxia[12]. A Veterans Affairs based observational cohort study examined 1233 consecutive patients referred for evaluation of OSA, of whom 544 were free from pre-existing diabetes. At a median follow-up time of 2.7 years, they found a multivariable-adjusted hazard ratio of 1.43 per quartile for incident diabetes[13]. In a case control study from Japan, Kono et al[14] found that OSA severity as assayed by apnea-hypopnea index (AHI) was a strong predictor of a number of metabolic syndrome parameters, such as hypertension, hyperglycemia and dyslipidemia, while body mass index and lowest arterial oxygen saturation during sleep did not. A British study involving patients with type 2 diabetes and the use of structured questionnaires and overnight oximetry found a 23% prevalence of OSA[15]. The Wisconsin sleep study found an increased incidence of diabetes in patients with OSAS, but the significance of OSAS disappeared after accounting for obesity[16]. In examining results from the Sleep Heart Health study involving 2656 subjects, Punjabi et al[9] found a relationship between OSA severity as measured by both AHI and degree of sleep-related oxygen desaturations and IR. More recently, the same investigators used an intravenous glucose tolerance test in a cohort of patients with sleep apnea and without diabetes mellitus and found that OSA, independent of adiposity, is associated with impairments in insulin sensitivity, glucose effectiveness and pancreatic β-cell function[17]. Similarly, in a study involving 270 subjects referred for polysomnography who did not have known diabetes mellitus, Ip et al[18] found that there was a strong association of OSA and insulin resistance in both obese and non-obese subjects and that both AHI and minimum oxygen saturation were independent determinants of insulin resistance.

PATHOPHYSIOLOGY

A variety of putative pathogenetic mechanisms have been described and studied that explores the interplay between intermittent hypoxia, sleep fragmentation and insulin resistance (Figure 1).

Figure 1.

Flow diagram demonstrating interplay between intermittent hypoxia and sleep fragmentation, and insulin resistance. OSA: Obstructive sleep apnea; NFκB: Nuclear factor kappa B; PA1: Plasminogen activator inhibitor 1; HPA: Hypothalamopituitary axis; IL: Interleukin; TNFα: Tumor necrosis factor-alpha; ISI: Insulin sensitivity index.

Hypoxia has been shown to induce a multitude of effects on adipocytes, including inflammatory activation, transcription of genes regulated by the hypoxia inducible factor-1 and endoplasmic reticulum stress[19,20]. Adiponectin is an important adipokine with protective effects against insulin resistance. At least two studies have examined the behavior of adipocytes exposed to in vitro intermittent hypoxia and have shown less adiponectin production, despite a significant upregulation of adiponectin mRNA expression[21,22]. Sleep disruption is an additional, potentially important mechanism by which OSA may affect metabolism[23]. In healthy subjects, sleep restriction was associated with IR, increased appetite and craving for carbohydrates[24,25]. Disruption of normal sleep architecture has also been shown to induce a pro-inflammatory state, with increased release of interleukin (IL)-6 and tumor necrosis factor (TNF)-α by circulating monocytes[26]. Sleep apnea has also been postulated to lead to dysregulation of the hypothalamo-pituitary axis and this may well play a central role in modulating insulin resistance and a predisposition to diabetes mellitus[27]. Nocturnal awakenings have been shown to be associated with pulsatile cortisol release[28] and autonomic activation.

EFFECT OF CPAP THERAPY ON IR

Continuous positive airway pressure (CPAP) has emerged as an effective therapy for OSA. In view of the evidence that OSA can lead to insulin resistance and abnormality in glucose metabolism, studies have been done by several investigators to see if CPAP therapy, in addition to eliminating apnea, hypopnea, desaturation and sympathetic surge in turn, can lead to improvement in insulin resistance and blood glucose control. We will undertake a review of the literature as it pertains to the effect of CPAP in blood sugar control and insulin resistance.

EFFECT OF CPAP ON GLUCOSE METABOLISM

A multitude of studies have explored the influence of CPAP therapy on blood sugar control. Many of these have had issues with study design, including the lack of a control group and small sample size. These studies are summarized in Table 1.

Table 1.

Effect of continuous positive airway pressure on blood glucose control

Study	Study design/ cohort	Sample size	Control group	Outcome/ measurements	Study duration	Conclusions	+/-
Stoohs et al[29]	OSA patients	5	None	Fasting glucose and insulin	2 mo	No change in either fasting or nocturnal insulin level	_
						Increase in nocturnal and fasting glucose
Saini et al[30]	OSA patients BMI 32.7 ± 2.3 Kg/m2	8	None	Glucose and insulin every 10 min interval during sleep	1 night	Mean insulin and glucose did not differ between pre treatment and treatment night	-
Davies et al[31]	OSA patients	10	Matched control	Fasting insulin, lipid profile	3 mo	No change in insulin level with CPAP	-
Brooks et al[32]	OSA patients with BMI 42.7 ± 4.3 kg/m2	10	None	Hyperinsulinemic euglycemic clamp	4 mo	Improvement in insulin responsiveness seen	+
Cooper et al[33]	OSA patients	6	None	Insulin and c-peptide sample every hour and glucose sample every 30 min during sleep	1 night	No changes in glucose, insulin and C-peptide with CPAP treatment	-
Saarelainen et al[34]	OSA patients	7	None	Hyperinsulinemic euglycemic clamp	3 mo	No change in insulin responsiveness	_
Pierzchala et al[35] (article in Polish)	Type 1 and type 2 diabetes patients with OSA	30	None	Blood glucose	6 mo	Better blood glucose control	+
Chin et al[36]	OSA patients	22	OSA patients	Oral glucose tolerance test with insulin measurement	6 mo	No change in glucose and insulin level except in patients who have lost weight	_
Ip et al[37]	OSA patients	30	Matched 30 non-OSA control	Fasting glucose and insulin	6 mo	No change in fasting glucose and insulin seen (decrease in Leptin and triglyceride was seen)	_
Smurra et al[38]	16 OSA patients; 10 from endocrine clinic and 6 other OSA patients	16	None	Oral glucose tolerance test in 10 patients and hyperinsulinemic euglycemic clamp in 6 patients	2 mo	No change in mean glycemia, insulin level or insulin responsiveness was seen	_
Harsch et al[39]	OSA patients	40	None	Hyperinsulinemic euglycemic clamp	3 mo	Improvement in insulin sensitivity at day 2 and 3 mo, in patient with BMI < 30, than in patients with BMI > 30	+
Harsch et al[40]	Type 2 diabetes patients with OSA	9	None	Hyperinsulinemic euglycemic clamp	3 mo	Insulin sensitivity was unchanged after 2 days, but significantly improved after 3 mo; glycemic control was unaffected after 3 mo	+/-
Babu et al[41]	Type 2 diabetes patients with OSA	25	None	HBA1c and post prandial blood glucose	3 mo	Decrease in HBA1c and postprandial am glucose level	+
Czupryniak et al[42]	Non diabetic OSA patients	9	None	Continuous glucose monitoring, plasma insulin, HOMA-IR	1 night	Mean blood glucose, fasting insulin and HOMA-IR were significantly higher with CPAP treatment	-
Hassaballa et al[43]	Type 2 diabetes and OSA (retrospective)	38	None	HBA1c	Approx. 3 mo	Decrease in HBA1c was seen with CPAP therapy	+
Lindberg et al[44]	OSA patients	28	Matched control without OSA	HOMA and fasting insulin	6 mo	Decrease in insulin resistance and fasting insulin	+
West et al[45]	Type 2 diabetes and OSA	42	Randomized, double blind	HOMA, hyperinsulinaemic euglycemic clamp, HBA1c, highly sensitive C-reactive protein	3 mo	No change in glycemic control or insulin resistance	_
Coughlin et al[46]	OSA patients	34	Randomized placebo-controlled blinded crossover trial	Insulin, fasting glucose, HOMA-IR	6 wk	No change in glucose or insulin resistance	_
Pallayova et al[47]	Type 2 diabetes with OSA	14	None	Continuous glucose monitoring	Several d	Reduction in nocturnal glucose variability and improved overnight glucose control	+
Wang et al[48]	Type 2 diabetes and OSA	30	None	HOMA	7 d	Improve ISI	+
Dawson et al[49]	Type 2 diabetes with OSA	20	None	Continuous glucose monitoring	Average 41 d (26-96 d)	Decrease in sleeping blood glucose seen	+
Steiropoulos et al[50]	Diabetes with OSA	56	None	HBA1c, fasting glucose, insulin level, HOMA-IR	6 mo	Only patients with CPAP use > 4 h/night showed decrease in HBA1c	+/-
Wei et al[51]	OSA patients	11	None	Fasting blood glucose, plasma insulin, HOMA-IR	4 d	Decrease in blood glucose and increase in insulin sensitivity seen	+
Oktay et al[52]	OSA and metabolic syndrome	20	None	Fasting blood glucose	1 yr	No difference in blood glucose seen	_
Lam et al[53]	OSA patients	61 (30 control and 31 study group)	Sham CPAP	Short insulin tolerance test	12 wk	Improvement in insulin sensitivity seen only in subjects with BMI ≥ 25	+/-
Garcia et al[54]	Obese OSA patients	20	None	OGTT, insulin level, Gherlin, adiponectin, leptin	6 mo	Increase insulin and IR; gherlin decrease, whereas leptin and adiponectin remains unchanged	-
Shpirer et al[55]	OSA patients	30	None	HBA1c	3-5 mo	Decrease in HBA1c in severe OSA patients	+

HOMA-IR: Homeostatic model assessment of insulin resistance; HBA1c: Glycosylated Hemoglobin; ISI: Insulin sensitivity index; OGTT: Oral glucose tolerance test; BMI: Body mass index; OSA: Obstructive sleep apnea; CPAP: Continuous positive airway pressure.

CONCLUSION

There is a strong association between OSA and diabetes mellitus. A multitude of pathophysiological perturbations have been demonstrated both in vitro and in vivo that demonstrate a close interrelationship, including inflammatory mediators of oxidative stress, as well as leptin resistance and hypothalamo-pituitary axis dysregulation. These effects are mediated secondary to both the effect of intermittent hypoxia as well as sleep fragmentation. Treatment with nightly CPAP leads to a resolution of both these behaviors and has been shown to be effective, not only in resolving daytime sleepiness, but also improving cardiovascular mortality. The data on the impact, if any, and its magnitude on glycemic control are neither convincing nor clear. The few randomized trials that have explored this impact have had mixed results, moreover confounded by the influence of morbid obesity as well as lack of optimal CPAP compliance. Future research is needed to clarify both the downstream mechanisms that stem from sleep fragmentation and the oxidative stress of intermittent hypoxia, as well as the impact of complete resolution of OSA with adequate CPAP compliance (adequately powered and randomized controlled design) on the metabolic profile of patients with OSA and diabetes mellitus in both obese and non-obese cohorts.