Snoring and witnessed sleep apnea is related to diabetes mellitus in women

1. Introduction 

The prevalence of diabetes mellitus is increasing and in global terms; approximately 30 million people had diabetes in 1985 and 177 million in 2000 [1]. Subjects with diabetes run an increased risk of cardiovascular disease, renal failure and retinopathy. Diabetes-related complications account for about 5.3% of the global causes of death [2].

Snoring is a sign of compromised upper airways and obstructive sleep apnea. The prevalence of habitual snoring among middle-aged men is 14–16%, while 6.5–6.7% of women snore habitually [3], [4], [5]. Many snorers suffer from daytime sleepiness and they run an increased risk of developing cardiovascular disease [6], [7], [8], [9].

Increasing interest has been shown in the association between snoring and sleep apnea and diabetes mellitus, glucose intolerance and insulin resistance [3], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22]. A relationship between snoring and diabetes mellitus independent of obesity has been shown in women [15], [21], [22] and to impaired glucose tolerance in non-obese men [16]. Most epidemiological studies regarding sleep-disordered breathing ask for snoring frequency, but do not include questions on witnessed sleep apneas.

We aimed to analyze the relationship between snoring, witnessed sleep apnea and diagnosed diabetes mellitus in an unselected population of men and women, and also to analyze possible gender related differences in such a relationship.

2. Methods 

Questions on sleep-disordered breathing were included in the Northern Sweden component of the WHO “Multinational monitoring of trends and determinants in cardiovascular disease” (MONICA) study in 1999 and 2004 [23]. The study consisted of a questionnaire relating to the risk factors for cardiovascular diseases and physical examinations. Participants were randomly selected from the population register in the two northernmost counties of Sweden and stratified for age into 10-year groups, between 25 and 79 years of age. A total of 10,756 men and women were invited, 8256 in 1999 and 2500 in 2004. The participants were given questionnaires and were invited for physical examinations.

The ethics committee at Umeå University Hospital approved the study protocols and all subjects gave their informed consent.

3. Results 

There were 7905 of 10,756 (73%) subjects who answered the questionnaires and attended visits for physical examinations. They were 53±13 years old, with a body mass index of 26.7±4.3kg/m2 and 4.3% had diagnosed diabetes mellitus. There were 4047 women and 3858 men. Men more often snored habitually than women (24.4% vs. 12.7%, respectively, p<0.001) and men more often reported witnessed sleep apneas (5.1% vs. 1.4%, respectively, p<0.001). Diabetes mellitus was more prevalent in men (5.0% vs. 3.6%, respectively, p=0.003). The frequencies of diabetes mellitus, snoring and witnessed sleep apnea in relation to age are presented in Table 1.

Table 1. Frequency of diabetes mellitus, habitual snoring and witnessed sleep apnea

	Age (years)	Numbers	Body mass indexa (kg/m2)	Diabetes mellitus (%)	Habitual snoring (%)	Witnessed sleep apnea (%)
	Women
	25–34	411	24.6±4.6	0.7	4.9	0.0
	35–44	780	25.5±4.6	1.5	11.3	1.6
	45–54	902	26.3±4.6	1.6	15.6	1.4
	55–64	953	27.2±4.7	4.4	14.7	1.8
	65–74	888	28.0±4.8	7.3	12.8	1.3
	75–79	113	27.5±4.8	9.3	8.8	1.1
	Total 25–79	4047	26.6±4.8	3.6	12.7	1.4
	Men
	25–34	368	25.8±3.8	0.5	20.3	1.2
	35–44	671	26.5±3.7	0.9	22.3	3.1
	45–54	891	27.1±3.6	2.3	31.2	6.7
	55–64	912	27.4±3.5	6.4	28.1	6.6
	65–74	890	27.2±3.8	9.2	18.5	5.6
	75–79	126	26.5±3.2	18.3	13.5	1.8
	Total 25–79	3858	26.9±3.7	5.0	24.4	5.1

a Mean±standard deviation.

Of snoring women, 6.5% suffered from diabetes mellitus vs. 3.0% in non-snoring women, p<0.001. Of snoring men, 4.8% suffered from diabetes mellitus vs. 4.9% in non-snoring men, p=0.9. Of women reporting witnessed sleep apnea, 11.8% had diabetes mellitus vs. 2.8% of remaining women, p=0.001. Of men reporting witnessed sleep apnea, 7.0% had diabetes mellitus vs. 4.7% of the remaining men, p=0.16.

Habitual snoring and witnessed sleep apnea were related to diabetes mellitus in women in the univariate analysis. Age, body mass index and waist circumference were other factors related to diabetes mellitus in both men and women (Table 2).

Table 2. Univariate logistic regression analysis

		Women			Men
		OR	95% CI	p-value	OR	95% CI	p-value
	Non-snoring	1			1
	Unknown snoring pattern	1.62	0.98–2.69	0.060	1.43	0.82–2.49	0.204
	Habitual snoring	2.25	1.49–3.38	<0.001	0.98	0.70–1.39	0.923
	No witnessed sleep apnea	1			1
	Unknown apnea pattern	1.85	1.27–2.70	0.001	1.25	0.87–1.80	0.224
	Witnessed sleep apnea	4.66	1.93–11.2	0.001	1.53	0.85–2.76	0.158
	Non-smoker	1			1
	Current smoking	0.62	0.38–1.02	0.062	0.72	0.44–1.19	0.201
	Age 25–44 (years)	1			1
	45–54	1.23	0.59–2.56	0.579	2.95	1.29–6.72	0.010
	55–64	3.61	1.99–6.55	<0.001	8.71	4.13–18.3	<0.001
	65–74	6.13	3.47–10.83	<0.001	13.0	6.26–27.0	<0.001
	75–79	7.73	3.39–17.65	<0.001	28.6	12.5–65.5	<0.001
	Body mass index (kg/m2)	1.11	1.08–1.14	<0.001	1.10	1.06–1.14	<0.001
	Waist circumference (cm)	1.06	1.04–1.07	<0.001	1.04	1.03–1.06	<0.001

Factors associated with diabetes mellitus. OR, odds ratio; CI, confidence interval.

In multiple regression analysis, habitual snoring remained significantly associated with diabetes mellitus among women with an adjusted OR=1.58 (95% CI 1.02–2.44, p=0.041) independent of smoking, age, body mass index and waist circumference, but not in men who had an adjusted OR=0.92 (95% CI 0.64–1.33, p=0.665) (Table 3). Witnessed sleep apnea was also independently related to diabetes mellitus in women, with an adjusted OR=3.29 (95% CI 1.20–8.32, p=0.012), but not in men with an adjusted OR=1.15 (95% CI 0.61–2.13, p=0.671) (Table 4).

Table 3. Multiple logistic regression analysis

		Women			Men
		OR	95% CI	p-value	OR	95% CI	p-value
	Non-snoring	1			1
	Unknown snoring pattern	1.03	0.61–1.75	0.909	1.11	0.63–1.99	0.714
	Habitual snoring	1.58	1.02–2.44	0.041	0.92	0.64–1.33	0.665
	Non-smoking	1			1
	Current smoking	0.76	0.45–1.29	0.312	0.79	0.47–1.33	0.371
	Age 25–44 (years)	1			1
	45–54	1.04	0.49–2.21	0.923	2.62	1.14–6.02	0.023
	55–64	2.89	1.56–5.37	0.001	7.45	3.53–15.8	<0.001
	65–74	4.68	2.58–8.49	<0.001	10.9	5.23–22.8	<0.001
	75–79	5.90	2.49–14.0	<0.001	26.0	11.2–60.1	<0.001
	Body mass index (kg/m2)	0.93	0.87–0.99	<0.001	1.00	0.94–1.08	0.910
	Waist circumference (cm)	1.08	1.05–1.11	<0.001	1.04	1.01–1.07	0.003

The association of habitual snoring and diabetes mellitus. OR, odds ratio; CI, confidence interval.

Table 4. Multiple logistic regression analysis

		Women			Men
		OR	95% CI	p-value	OR	95% CI	p-value
	No witnessed sleep apnea	1			1
	Unknown apnea pattern	1.29	0.87–1.92	0.207	1.11	0.76–1.62	0.574
	Witnessed sleep apnea	3.29	1.20–8.32	0.012	1.15	0.61–2.13	0.671
	Non-smoking	1		1
	Current smoking	0.85	0.50–1.45	0.553	0.85	0.50–1.42	0.526
	Age 25–44 (years)	1			1
	45–54	1.23	0.52–2.44	0.758	2.57	1.12–5.90	0.026
	55–64	2.72	1.43–5.20	0.002	6.69	3.15–14.2	<0.001
	65–74	4.86	2.61–9.05	<0.001	11.0	5.26–22.9	<0.001
	75–79	4.72	2.23–14.6	<0.001	23.7	10.1–55.8	<0.001
	Body mass index (kg/m2)	0.93	0.86–0.99	<0.041	1.01	0.94–1.08	0.858
	Waist circumference (cm)	1.08	1.05–1.11	<0.001	1.04	1.01–1.06	0.004

The association of witnessed sleep apnea and diabetes mellitus. OR, odds ratio; CI, confidence interval.

A sub-analysis was done in the younger half of men and women aged 25–54 years old, including 28 men and 29 women with diabetes mellitus. The adjusted OR for habitual snoring and diabetes in this age group was 1.61 (95% CI 0.63–4.13, p=0.321) in women, and 1.53 (95% CI 0.69–3.38, p=0.296) in men. The adjusted OR for witnessed sleep apnea and diabetes mellitus was 4.10 (95% CI 0.87–19.3, p=0.074) in women, and 3.84 (95% CI 1.36–10.9, p=0.011) in men.

4. Discussion 

Women in the present study aged 25–79 years old reporting that they snored always or almost always had a 58% increased frequency of diabetes mellitus, and women reporting witnessed sleep apnea had an over threefold increased frequency of diabetes mellitus independent of smoking, age, body mass index and waist circumference. Neither snoring nor witnessed sleep apnea was related to diabetes mellitus in men, except for witnessed sleep apnea in men aged 25–54 years old. These men had an almost fourfold increased frequency of diabetes mellitus independent of possible confounders.

Al-Delaimy and colleagues reported that snoring women ran a twofold higher risk of developing diabetes mellitus compared with non-snoring women [21]. Another prospective study focusing exclusively on men, reported that obese snorers were at an increased risk of developing diabetes mellitus compared to obese non-snorers, but snoring per se was not a significant predictor of developing diabetes mellitus after adjustments for confounders [3]. Two cross-sectional studies included both men and women [10], [15]. One reported an association between habitual snoring and diabetes mellitus in 593 men and women [15], while the other only found an association between snoring and diabetes mellitus in older women [10]. A recent study by Reichmuth and colleagues found that sleep apnea was independently related to diabetes mellitus in a cross-sectional analysis of 1387 subjects (44% women) but not in the prospective analysis [20]. The gender difference observed in our study could explain why snoring was related to diabetes mellitus in some studies but not in other studies.

Men have lower life expectancy and higher cardiovascular risk than women and we cannot rule out that the more affected men, in terms of diabetes mellitus and snoring or sleep apnea, have died, perhaps because of this relationship. The finding that witnessed sleep apnea was independently related to diabetes mellitus in men aged 25–55 years old but not for the whole sample of men aged 25–79 years old is most interesting and in support of this hypothesis.

Both diabetes mellitus and sleep disordered breathing are influenced by sexual hormones. Sleep apnea increases after menopause, and can be reduced or alleviated with hormone replacement therapy [24], [25], [26], [27]. Both snoring and diabetes increase during pregnancy [9], [28]. Polycystic ovary syndrome is a common condition and occurs in about 7% of premenopausal women [29]. Women with polycystic ovary syndrome suffer from hyperandrogenism and chronic anovulation with infertility and hirsutism as a result. These women often develop diabetes mellitus type 2, and they often suffer from sleep apnea [30], [31]. It is possible that some of the observed gender differences in the relationship of snoring, witnessed sleep apnea and diabetes mellitus could be explained by a co-existing or previous polycystic ovary syndrome.

In support that nocturnal upper airway obstructions may contribute to the development of diabetes mellitus, snoring and especially witnessed sleep apneas were also related to diabetes mellitus in the present subjects, including unselected men and women from the population. Increased sympathetic activation from sleep apneas [32], [33], the release of pro-inflammatory cytokines from cyclic hypoxia and increased evening serum cortisol levels from sleep deprivation have been suggested as a possible explanation linking obstructive sleep apnea to diabetes mellitus [3], [11], [12], [14], [17], [19], [21]. From the present study design we can only speculate about causes and consequences. It is also possible that snoring and sleep apneas are due to polyneuropathy related to diabetes mellitus. Other limitations of the study are the lack of data about type, severity and duration of diabetes mellitus.

Both diabetes mellitus and sleep apnea are related to cardiovascular diseases and early death. We observed that snoring and possibly sleep apnea are increased in women and younger men with diabetes mellitus. Our results also raise the question of whether men with diabetes and sleep apnea run a particular risk for early death. Harsch and colleagues reported improved insulin sensitivity with continuous positive airway pressure among patients with diabetes mellitus type 2 and among non-diabetic sleep apnea patients [19], [34]. It is possible that the treatment of sleep apnea in men and women with diabetes mellitus could serve as a complement to insulin therapy and oral antidiabetic medication.

In conclusion, snoring and especially witnessed sleep apnea are related to diabetes mellitus in women. Witnessed sleep apnea is related to diabetes mellitus in men younger than 55 years old.