Joint effects of self-reported sleep and modifiable physical activity on risk of dyslipidaemia in women aged 45-55 years: a cross-sectional study.

OBJECTIVES: Modifiable physical activity (PA) plays an important role in dyslipidaemia risk in middle-aged women with sleep problems, especially perimenopausal women. We aimed to explore the joint effects of sleep and PA on the risk of dyslipidaemia in women aged 45-55 years, and the extent to which PA moderated the effect of sleep on the risk of dyslipidaemia.
DESIGN: A cross-sectional study.
SETTING: This study was based on the survey of Chronic Disease and Nutrition Monitoring in Adults in Inner Mongolia in 2015. PARTICIPANTS: 721 women aged 45-55 years were included. OUTCOME MEASUREMENT: PA was measured by the Global Physical Activity Questionnaire. Sleep was measured by questionnaire formulated by the Chinese Center for Disease Control and Prevention. Multivariate logistic regression analyses were performed to determine the joint effects of sleep and PA on dyslipidaemia risk. OR and 95% CI were reported.
RESULTS: Among all participants, 60.6% had sleep problems, 29.0% had low PA and 41.1% had dyslipidaemia. Women with sleep problems had higher dyslipidaemia risk than women without sleep problems, irrespective of low, moderate or high PA, with OR (95% CI) of 4.24 (2.40 to 7.49), 3.14 (1.80 to 5.49) and 2.04 (1.20 to 3.48), respectively. PA could not completely attenuate the negative association between sleep and dyslipidaemia risk. With PA increased from low to high, the OR of dyslipidaemia decreased by 2.20. Women with sleep problems and low PA had higher risks of high total cholesterol, high triglyceride, low high-density lipoprotein cholesterol and high low-density lipoprotein cholesterol than women without sleep problems and high PA, with OR (95% CI) of 2.51 (1.18 to 5.35), 2.42 (1.23 to 4.74), 2.88 (1.44 to 5.74) and 2.52 (1.12 to 5.70), respectively.
CONCLUSIONS: Among women aged 45-55 years, the joint effects of self-reported sleep and PA on dyslipidaemia risk were more marked for sleep than for PA. Modifiable PA is a widely accessible and effective intervention to reduce the dyslipidaemia risk in women with sleep problems, particularly among perimenopausal women. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

This study was based on the survey of Chronic Disease and Nutrition Monitoring in Adults, covered urban, agricultural and pastoral areas in Inner Mongolia. The participants were recruited consecutively used multistage stratified cluster random sampling, which were generally representative.

The study explored the joint effects of sleep and physical activity (PA) on the risk of dyslipidaemia in women aged 45–55 years, and the extent to which PA moderated the effect of sleep on the risk of dyslipidaemia.

Due to the cross-sectional design of the study, associations between self-reported sleep, PA and the risk of dyslipidaemia can be inferred but not causality.

Introduction

Dyslipidaemia is a major risk factor for cardiovascular disease.1 2 Elevated low-density lipoprotein cholesterol (LDL-C) accounted for 4.32 million deaths and 94.9 million disability-adjusted life years worldwide in 2017.3 Furthermore, elevated total cholesterol (TC) and triglyceride (TG) can increase the risk of type 2 diabetes mellitus,4 and reduced high-density lipoprotein cholesterol (HDL-C) can worsen the outcomes of patients with coronary heart disease.5

Dyslipidaemia includes elevated TC, elevated TG, elevated LDL-C and low HDL-C.6 Risk factors of dyslipidaemia includes age,7 low oestrogen levels,8 sleep disorder,9 smoking10 and insufficient aerobic exercise.11 Among perimenopausal women, aged 45–55 years,12 13 with decreasing oestrogen levels, the incidence of dyslipidaemia has increased significantly.14 15 The prevalence of dyslipidaemia was reported to be 38.7% among middle-aged and elderly women in China.16 Thus, identifying modifiable risk factors for dyslipidaemia is very important for public health.

Sleep problems tend to become more common in middle-aged and elderly women, and include difficulty in falling asleep, early awakening and awakening during the night.17–20 Poor sleep quality was significantly associated with an increased risk of dyslipidaemia.21 22 Poor sleep quality was also correlated with high TG and TC,23 but negatively correlated with HDL-C.24 A potential mechanism for the associations may be that women in the perimenopausal period, had gradually decreasing hormone levels when their sleep problems occurred, leading to dyslipidaemia.25

Physical activity (PA) is a modifiable and conventional influencing factor for dyslipidaemia. Several studies demonstrated that low PA had adverse effects on prevalence of dyslipidaemia.16 26 HDL-C was found to increase after PA intervention.27 PA led to high energy consumption, which reduced the level of TG.28 For patients with dyslipidaemia, increased PA significantly reduced their risk of death.29

There is some evidence that the association between sleep and PA is bidirectional. Moderate-to-vigorous intensity PA was associated with better sleep quality.30 31 Poor sleep quality limited PA,32 and sleep problems were associated with insufficient PA.33 Therefore, the mechanism for the effects of sleep and PA on dyslipidaemia is very complex.

PA is a modifiable lifestyle factor, but sleep problems are important symptoms of neurological diseases that are difficult to regulate.34Casas et al35 reported significant linear trends in TG and HDL-C among leisure-time activity and sleep quality categories in middle-aged women. However, little is known about the associations of PA, including occupational, transition and leisure-time activity, and sleep with risk of dyslipidaemia in women aged 45–55 years. This study aimed to explore the joint effects of sleep and PA on the risk of dyslipidaemia in women aged 45–55 years, and the extent to which PA moderated the effect of sleep on the risk of dyslipidaemia to provide some effective suggestions for the management of dyslipidaemia in perimenopausal women.

Methods

Design and setting of the study

The data for this cross-sectional study were collected from the survey of Chronic Disease and Nutrition Monitoring in Adults in Inner Mongolia in 2015, conducted by the Chinese Center for Disease Control and Prevention (CDC). The survey was conducted across eight monitoring sites, covered urban, agricultural and pastoral areas in Inner Mongolia. Participants, aged ≥18 years were selected by multistage-stratified cluster random sampling. All participants received on-site investigations that included questionnaires, anthropometric measurements and fasting blood sample collections.

Women aged 45–55 years were enrolled in the present study (n=821). Subsequently, 100 women were excluded for missing weight, height or waist circumference data (n=25), missing serum lipid testing data (n=8), missing alcohol drinking data (n=1), reporting that they did not drink but indicating alcohol intake (n=32) and having mean total PA time exceeding 1440 min/day (n=34). Finally, a total of 721 women were included in the study.

Data collection

The questionnaire used in the study was from Chinese CDC and evaluated by experts, with high reliability and validity. General information including age, ethnicity, educational level, marital status and residence, lifestyle variables including smoking status, alcohol drinking status, sleep and PA, and history of dyslipidaemia were obtained by trained investigators through face-to-face interviews.

Anthropometric data including height, weight and waist circumference were measured by trained investigators. Fasting venous blood samples were collected by trained laboratory staff members.36

Measurements

Dyslipidaemia definition

Dyslipidaemia was defined as having any one of the following: high TC (serum TC ≥240 mg/dL), high TG (serum TG ≥200 mg/dL), high LDL-C (serum LDL-C ≥160 mg/dL), low HDL-C (serum HDL-C <40 mg/dL) and self-reported dyslipidaemia diagnosis.37

Sleep assessment

Self-reported sleep was measured by the following questions: during the past 30 days, for more than 3 days per week (including 3 days), did you have any one of the following problems: (1) snoring, asphyxiating or suffocating; (2) difficulty in falling asleep (more than 30 min); (3) midway wakefulness twice or more; (4) medication for at least 1 day to help sleep; and (5) waking up early and difficulty in falling asleep again. Participants could respond ‘yes’ or ‘no’. We defined sleep problems as at least one answer of ‘yes’. These similar questions of sleep assessment have used in previous studies.17 38

Assessment of PA

PA was collected using the Global Physical Activity Questionnaire (GPAQ).39 Different domain-specific PAs (work, transportation and recreational activities) were classified into moderate and vigorous intensity. For assessment of total PA, we calculated the metabolic equivalent (MET) minutes per week according to the GPAQ analysis guide40: MET-min/week = (moderate minutes×4.0 METs) + (vigorous minutes×8.0 METs). Total PA was categorised as low (0–1908 MET-min/week), moderate (1909–5129 MET-min/week) and high (≥5130 MET-min/week).41

Covariates

Covariates included age, ethnicity (Han/Mongolian/other minority), educational level (primary school or lower/junior high school/senior high school and above), marital status (single/married/others) and residence (rural/urban). Lifestyle variables included smoking status (yes/no) and alcohol drinking status (yes/no). Anthropometric data included height, weight and waist circumference (<85 cm, ≥85 cm). Body mass index was calculated as weight (kg)/height2 (m2) and categorised as normal or underweight (<24.0 kg/m2), overweight (24.0–27.9 kg/m2) or obese (≥28.0 kg/m2).42

Statistical analysis

Continuous variables were shown as mean±SD. Categorical variables were expressed as rate or proportion. The t-test, analysis of variance and χ2 test were used to analyse differences in sleep, PA and dyslipidaemia in the participants with different characteristics. Multivariate logistic regression analyses were conducted to examine the associations of sleep and PA with risk of dyslipidaemia. The results were reported as OR with 95% CI. We defined α=0.05 as the significance level, and two-tailed p<0.05 was considered to indicate statistical significance. All data analyses were performed using IBM SPSS Statistics V.19.0 (IBM Corp).

Patient and public involvement

This research was done without patient involvement. Patients were not invited to comment on the study design and were not consulted to develop patient-relevant outcomes or interpret the results. Patients were not invited to contribute to the writing or editing of this document for readability or accuracy.

Results

Demographic characteristics of the participants

Seven hundred and twenty-one women were included in this study and the mean age was 50.2 years (SD=2.8 years). Among all participants, 593 (82.2%) were Han ethnicity, 430 (61.0%) were from rural areas, 317 (44.0%) had primary school or lower educational level and the majority were married (96.8%). Six hundred and eleven (84.7%) were non-drinkers, 675 (93.6%) were non-smokers, 237 (32.9%) were normal or underweight and 425 (58.9%) had waist circumference <85 cm.

Four hundred and thirty-seven (60.6%) had sleep problems and 209 (29.0%) had low PA. There were significant differences in sleep among people with different age, smoking status, alcohol drinking status and residence status, and significant differences in PA among people with different educational level and residence (all p<0.05, table 1). There was no significant difference in PA between the different sleep groups (p>0.05, table 1).

Table 1

Characteristics of sleep and physical activity in women

	Sleep problems		P value	Physical activity			P value
	No (n=284)	Yes (n=437)		Low (n=209)	Moderate (n=215)	High (n=297)
Age (years)	49.8 (2.8)	50.5 (2.8)	0.001**	50.1 (2.9)	50.5 (2.9)	50.1 (2.8)	0.285
Ethnicity
Han	230 (38.8)	363 (61.2)	0.757	169 (28.5)	173 (29.2)	251 (42.3)	0.423
Mongolian	37 (41.6)	52 (58.4)		28 (31.5)	26 (29.2)	35 (39.3)
Other minority	17 (43.6)	22 (56.4)		12 (30.8)	16 (41.0)	11 (28.2)
Educational level
Primary school or lower	127 (40.1)	190 (59.9)	0.068	75 (23.7)	87 (27.4)	155 (48.9)	0.002**
Junior high school	87 (34.5)	165 (65.5)		79 (31.4)	78 (31.0)	95 (37.6)
Senior high school and above	70 (46.1)	82 (53.9)		55 (36.2)	50 (32.9)	47 (30.9)
Marital status
Singled	2 (66.7)	1 (33.3)	0.578	1 (33.3)	2 (66.7)	0 (0)	0.097
Married	275 (39.4)	423 (60.6)		205 (29.4)	203 (29.1)	290 (41.5)
Others	7 (35.0)	13 (65.0)		3 (15.0)	10 (50.0)	7 (35.0)
Smoking status
No	273 (40.4)	402 (59.6)	0.026*	197 (29.2)	204 (30.2)	274 (40.6)	0.440
Yes	11 (23.9)	35 (76.1)		12 (26.1)	11 (23.9)	23 (50.0)
Alcohol drinking status
No	253 (41.4)	358 (58.6)	0.009**	185 (30.3)	180 (29.5)	246 (40.3)	0.190
Yes	31 (28.2)	79 (71.8)		24 (21.8)	35 (31.8)	51 (46.4)
Residence
Rural	188 (42.7)	252 (57.3)	0.022*	121 (27.5)	120 (27.3)	199 (45.2)	0.020*
Urban	96 (34.2)	185 (65.8)		88 (31.3)	95 (33.8)	98 (34.9)
Body mass index
Normal or underweight	94 (39.7)	143 (60.3)	0.723	69 (29.1)	67 (28.3)	101 (42.6)	0.966
Overweight	132 (40.5)	194 (59.5)		93 (28.5)	101 (31.0)	132 (40.5)
Obese	58 (36.7)	100 (63.3)		47 (29.8)	47 (29.8)	64 (40.4)
Waist circumference (cm)
<85	163 (38.4)	262 (61.6)	0.495	119 (28.0)	132 (31.1)	174 (40.9)	0.639
≥85	121 (40.9)	175 (59.1)		90 (30.4)	83 (28.0)	123 (41.6)
Sleep problems
No	—	—		83 (29.2)	82 (28.9)	119 (41.9)	0.902
Yes	—	—		126 (28.8)	133 (30.4)	178 (40.7)

*p<0.05, **p<0.01.

Prevalence of dyslipidaemia and its components

The prevalence of dyslipidaemia in women with sleep problems was significantly higher than that of women without sleep problems (47.4% vs 31.3%; p<0.001). Women with sleep problems also had higher prevalence of high TC, high TG than women without sleep problems (14.2% vs7.7%, 23.3% vs 15.8%; all p≤0.05). Women with low PA had a higher prevalence of dyslipidaemia (49.8% vs 43.7% vs 33.0%; p=0.001), and low HDL-C (27.7% vs 23.3% vs 18.5%; p=0.049) compared with moderate or high PA. (table 2).

Table 2

Prevalence of dyslipidaemia and its components in women

	All	Sleep problems		P value	Physical activity			P value
		No	Yes		Low	Moderate	High
Dyslipidaemia
No	425 (58.9)	195 (68.7)	230 (52.6)	<0.001***	105 (50.2)	121 (56.3)	199 (67.0)	0.001**
Yes	296 (41.1)	89 (31.3)	207 (47.4)		104 (49.8)	94 (43.7)	98 (33.0)
High TC
No	673 (88.3)	262 (92.3)	375 (85.8)	0.008**	177 (84.7)	188 (87.4)	272 (91.5)	0.052
Yes	84 (11.7)	22 (7.7)	62 (14.2)		32 (15.3)	27 (12.6)	25 (8.5)
High TG
No	574 (79.6)	239 (84.2)	335 (76.7)	0.015*	160 (76.7)	168 (78.1)	246 (82.8)	0.184
Yes	147 (20.4)	45 (15.8)	102 (23.3)		49 (23.3)	47 (21.9)	51 (17.2)
Low HDL-C
No	558 (77.4)	229 (80.6)	329 (75.3)	0.093	151 (72.3)	165 (76.7)	242 (81.5)	0.049*
Yes	163 (22.6)	55 (19.4)	108 (24.7)		58 (27.7)	50 (23.3)	55 (18.5)
High LDL-C
No	655 (90.8)	264 (93.0)	391 (89.5)	0.113	183 (87.6)	196 (91.2)	276 (92.9)	0.117
Yes	66 (9.2)	20 (7.0)	46 (10.5)		26 (12.4)	19 (8.8)	21 (7.1)

*p<0.05, **p<0.01, ***p<0.001.

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride.

Independent effect of PA and sleep on dyslipidaemia and its components

In the multivariate logistic analyses, the effects of PA and sleep on risk of dyslipidaemia remained evident. Compared with women without sleep problems, the adjusted OR (95% CI) for dyslipidaemia, high TC and high TG in women with sleep problems were 2.15 (1.54 to 2.99), 1.81 (1.08 to 3.04) and 1.64 (1.10 to 2.46), respectively. Furthermore, compared with women with high PA, the adjusted OR (95% CI) for dyslipidaemia in women with low PA was 1.94 (1.32 to 2.84) (table 3).

Table 3

Independent effect of physical activity and sleep on dyslipidaemia and its components

	Model 1†	P value	Model 2‡	P value
	OR (95% CI)		OR (95% CI)
Dyslipidaemia
Physical activity
High	1.00 (reference)	<0.001***	1.00 (reference)	0.002**
Moderate	1.58 (1.10 to 2.27)		1.55 (1.06 to 2.27)
Low	2.01 (1.40 to 2.89)		1.94 (1.32 to 2.84)
Sleep problems
No	1.00 (reference)	<0.001***	1.00 (reference)	<0.001***
Yes	1.97 (1.44 to 2.70)		2.15 (1.54 to 2.99)
High TC
Physical activity
High	1.00 (reference)	0.017*	1.00 (reference)	0.059
Moderate	1.56 (0.88 to 2.78)		1.48 (0.83 to 2.65)
Low	1.97 (1.13 to 3.43)		1.98 (1.13 to 3.47)
Sleep problems
No	1.00 (reference)	0.009**	1.00 (reference)	0.025*
Yes	1.97 (1.18 to 3.28)		1.81 (1.08 to 3.04)
High TG
Physical activity
High	1.00 (reference)	0.076	1.00 (reference)	0.348
Moderate	1.35 (0.88 to 2.10)		1.30 (0.82 to 2.06)
Low	1.48 (0.95 to 2.29)		1.37 (0.87 to 2.17)
Sleep problems
No	1.00 (reference)	0.015*	1.00 (reference)	0.015*
Yes	1.62 (1.10 to 2.38)		1.64 (1.10 to 2.46)
Low HDL-C
Physical activity
High	1.00 (reference)	0.014*	1.00 (reference)	0.083
Moderate	1.33 (0.87 to 2.05)		1.32 (0.84 to 2.07)
Low	1.69 (1.11 to 2.58)		1.65 (1.06 to 2.57)
Sleep problems
No	1.00 (reference)	0.094	1.00 (reference)	0.073
Yes	1.37 (0.95 to 1.97)		1.43 (0.97 to 2.10)
High LDL-C
Physical activity
High	1.00 (reference)	0.043*	1.00 (reference)	0.092
Moderate	1.27 (0.67 to 2.43)		1.38 (0.72 to 2.65)
Low	1.87 (1.02 to 3.42)		1.98 (1.07 to 3.65)
Sleep problems
No	1.00 (reference)	0.115	1.00 (reference)	0.064
Yes	1.55 (0.90 to 2.69)		1.69 (0.97 to 2.95)

*p<0.05, **p<0.01, ***p<0.001.

†Model 1: Unadjusted.

‡Model 2: Adjusted for age, ethnicity, education level, marital status, residence, smoking status, alcohol drinking status, waist circumference and physical activity (association of sleep problems) or sleep problems (association of physical activity).

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride.

Joint effects of sleep and PA on the risk of dyslipidaemia and its components

The participants were categorised into six subgroups according to their joint PA and sleep characteristics. After adjustment for confounders, the joint effects of PA and sleep on dyslipidaemia were significant. Women with sleep problems had higher risk of dyslipidaemia than women without sleep problems, irrespective of low, moderate or high PA, with OR (95% CI) of 4.24 (2.40 to 7.49), 3.14 (1.80 to 5.49) and 2.04 (1.20 to 3.48), respectively, (figure 1). PA could not completely attenuate the negative association between sleep problems and risk of dyslipidaemia. With PA increased from low to high, the OR of dyslipidaemia decreased by 2.20 (table 4).

Figure 1

The simple effect of sleep with dyslipidaemia. PA, physical activity.

Table 4

The ΔOR of the dyslipidaemia and its components in women with sleep problems

	Sleep problems	Physical activity			ΔOR
		Low OR (95% CI)	Moderate OR (95% CI)	High OR (95% CI)
Dyslipidaemia	Yes	4.24 (2.40 to 7.49)*	3.14 (1.80 to 5.49)*	2.04 (1.20 to 3.48)*	2.20
	No	1.73 (0.92 to 3.24)	1.58 (0.83 to 2.98)	1.00 (reference)
High TC	Yes	2.51 (1.18 to 5.35)*	1.65 (0.75 to 3.60)	0.79 (0.34 to 1.83)	1.72
	No	0.65 (0.22 to 1.94)	0.76 (0.27 to 2.16)	1.00 (reference)
High TG	Yes	2.42 (1.23 to 4.74)*	1.87 (0.94 to 3.71)	1.56 (0.80 to 3.01)	0.86
	No	1.05 (0.47 to 2.37)	1.49 (0.68 to 3.24)	1.00 (reference)
Low HDL-C	Yes	2.88 (1.44 to 5.74)*	2.08 (1.03 to 4.20)*	1.97 (1.01 to 3.84)*	0.91
	No	2.12 (1.00 to 4.49)	2.00 (0.92 to 4.31)	1.00 (reference)
High LDL-C	Yes	2.52 (1.12 to 5.70)*	1.52 (0.64 to 3.62)	0.78 (0.32 to 1.92)	1.74
	No	0.74 (0.24 to 2.25)	0.78 (0.25 to 2.38)	1.00 (reference)

*p<0.05.

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride.

Likewise, women with sleep problems and low PA had higher risks of high TC, high TG, low HDL-C and high LDL-C than women without sleep problems and high PA, with OR (95% CI) of 2.51 (1.18 to 5.35), 2.42 (1.23 to 4.74), 2.88 (1.44 to 5.74) and 2.52 (1.12 to 5.70), respectively. With PA increased from low to high, the OR of low HDL-C decreased by 0.91 (table 4). Sleep problems were associated with increased risks of dyslipidaemia (OR=3.14; 95% CI: 1.80 to 5.49) and low HDL-C (OR=2.08; 95% CI: 1.03 to 4.20) among women with moderate PA. In addition, high PA was associated with increased risk of low HDL-C (OR=1.97; 95% CI: 1.01 to 3.84) among women with sleep problems (figure 2A–D).

Figure 2

The simple effect of sleep with the components of dyslipidaemia; (A) TC, total cholesterol; (B) TG, triglyceride; (C) HDL-C, high-density lipoprotein cholesterol; (D) LDL-C, low-density lipoprotein cholesterol. PA, physical activity.

Discussion

In the present study, the prevalence of dyslipidaemia was 41.1% in women aged 45–55 years, 60.6% of participants had sleep problems and 29.0% participants had low PA. The prevalence of dyslipidaemia was 47.4% in women with sleep problems, and significantly higher than that in women without sleep problems.

Low PA and sleep problems were interactively associated with an increased risk of dyslipidaemia in women aged 45–55 years. Furthermore, the effects on dyslipidaemia were more marked for sleep than for PA. We found that PA attenuated but did not eliminate the excess risk of dyslipidaemia associated with sleep problems. With PA increased from low to high, the OR of dyslipidaemia and low HDL-C decreased by 2.20 and 0.91, respectively.

Comparisons with other studies

Among middle-aged and older people, PA had significant positive associations with several health conditions, including reduced risks of cardiovascular disease,43 diabetes44 and hypertension.45 Our study showed that high PA was associated with lower risk of dyslipidaemia and its components, included high TC and high TG in women aged 45–55 years. In our study, self-reported sleep was measured by a questionnaire that contained all dimensions of the widely employed Pittsburgh Sleep Quality Index (PSQI), and clearly distinguished whether women had sleep problems. Consistent with our results, previous studies reported that sleep problems measured by the PSQI were associated with the increased risk of dyslipidaemia and its components.23 24

To date, epidemiological evidence regarding the joint effects of PA and sleep on dyslipidaemia has been limited. Casas et al35 reported that low leisure-time activity was negatively associated with high TC and low HDL-C. The study pointed out that the joint effects on serum lipids were more marked for leisure-time activity than for sleep quality. Contrary to their results, the effect on risk of dyslipidaemia were more marked for sleep than for PA in the present study. These discrepancies among the findings may be partly explained by differences in the measurement methods for PA and sleep quality. In our study, we measured PA by the GPAQ, which included occupational, transition and leisure-time PA, and used a questionnaire to assess sleep quality. Whereas Casas, et al only analysed the leisure-time PA, and measured sleep quality by PSQI. What is more, the women participants were aged 45–55 years, who were in the perimenopausal period accompanied by changes in hormone levels. Previous study showed that daily PA and energy expenditure significantly decreased during the perimenopausal period.46 Santoro showed that poor sleep quality became more common in perimenopausal women associated with the hormonal changes.47 Furthermore, with decreasing oestrogen levels, the risk of dyslipidaemia of perimenopausal women has increased significantly.15

There are several potential factors to address for the difference between the simple effects of sleep and PA on the risk of dyslipidaemia. Age was positively associated with risk of dyslipidaemia.9 48 In our study, women with sleep problems were older, and women with dyslipidaemia were also older. Alcohol drinking has been suggested as a possible risk factor for dyslipidaemia.49 50 In our study, women with sleep problems were more likely to drink alcohol, which may be another reason why sleep problems had a greater effect on dyslipidaemia.

Our results were consistent with previous studies regarding the interactive associations of sleep and PA with chronic pain,51 somatic symptoms,52 insulin resistance53 and depressive symptoms.54 All of these studies highlighted the health benefits of PA and suggested that any adverse effects of sleep problems on those outcomes may be reduced by regular PA. In our study, with PA increased from low to high, the OR of dyslipidaemia decreased by 2.20. Meanwhile, contrary to other studies,32 33 there was no significant association between PA and sleep problems, suggesting that increasing PA may not improve sleep quality in middle-aged women.

Modifiable PA has an effective protection on human’s health against chronic diseases55–57 especially in participants with sleep problems, even if PA did not completely counteract the risk among sleep problems individuals. Modifiable PA are widely accessible and effective intervention to reduce the risk of dyslipidaemia and other chronic diseases in women with sleep problems, especially in perimenopausal women.

Strengths and limitations

Our study has some strengths and limitations. To our knowledge, there were relatively few studies on the joint effects of sleep and PA on risk of dyslipidaemia. This study was based on a survey of chronic disease and nutrition in Chinese adults, covered urban, agricultural and pastoral areas in Inner Mongolia. The participants were recruited consecutively used multistage cluster random sampling, which were generally representative. However, our study was a cross-sectional study, which only showed the relationship between sleep, PA and the risk of dyslipidaemia, not the causal association. In our study, self-reported sleep was measured by a questionnaire that included five questions. Although, these five questions consist all dimensions of the widely employed PSQI, there were some biases in our results due to no special scale used. Therefore, it is necessary to conduct more prospective and experimental research on the joint effects of sleep and PA on the risk of dyslipidaemia in women in the perimenopausal period.

Conclusions

In women aged 45–55 years, the joint effects of self-reported sleep and PA on risk of dyslipidaemia were more marked for sleep than for PA. Women aged 45–55 years with sleep problems and low PA should be given more attention. Modifiable PA is a widely accessible and effective intervention to reduce the risk of dyslipidaemia in women with sleep problems, particularly among women during the perimenopausal period. The study can provide some effective suggestions for the management of dyslipidaemia in perimenopausal women. Further research is warranted to identify available services for perimenopausal women to improve their quality of life.