Risk Factors of Sleep Disorder After Stroke:

Risk factors of sleep disorder after stroke: A Meta analysis

Running title: Risk factors of sleep disorder after stroke

Highlights:

  1. Diabetes mellitus has a 41% increased risk of sleep disorder in stroke patients
  2. Alcohol use has a 59% increased risk of sleep disorder in stroke patients
  3. Habitual snoring has a 1377% increased risk of sleep disorder in stroke patients
  4. Body-mass index has a 17% decreased risk of sleep disorder in stroke patients

Abstract

Objectives: To explore the related risk factors of sleep disorder in stroke patients by conducting a meta-analysis

Methods: PubMed, Medline, Springer, Elsevier Science Direct, Cochrane Library and Google scholar database were searched from inception up to May 2014. Studies investigating risk factors for sleep disorder after stroke were included. Characteristics including author’s name, year of publication, country, sample size, age of participants and gender were extracted independently by two reviewers.

Results: A total of 8 studies, involving 1381 patients (578 experimental groups and 803 control groups) were eligible for the meta-analysis. The meta-analysis showed that there were significant association between risk of sleep disorder after stoke between diabetes mellitus (OR = 1.41, 95% CI = 1.09 to 1.84, P < 0.05), alcohol use (OR = 1.59, 95% CI = 1.19 to 2.12, P < 0.05), habitual snoring (OR = 14.77, 95% CI = 5.52 to 39.53, P < 0.05), and body-mass index (WMD = 0.83, 95% CI = 0.63 to 1.03, P < 0.05) when compared with control groups, yet with no significant between-study heterogeneity. In addition, there were no significant association between risk of sleep disorder after stoke and gender (OR = 1.23, 95% CI = 0.72 to 2.1, P > 0.05), hypertension (OR = 1.22, 95% CI = 0.83 to 1.80, P > 0.05), dyslipidemia (OR = 0.96, 95% CI = 0.70 to 1.33, P > 0.05), smoking (OR = 1.27, 95% CI = 0.73 to 2.20, P > 0.05), previous stroke (OR = 1.05, 95% CI = 0.74 to 1.49, P > 0.05), chronic obstructive pulmonary disease (OR = 0.79, 95% CI = 0.40 to 1.57, P > 0.05) when compared with control groups. Furthermore, the results of Egger test showed no statistical publication bias.

Conclusions: Diabetes mellitus, alcohol use, habitual snoring and body-mass index are associated with risk of sleep disorder in patients with stroke.

Key words: Sleep disorder; Risk factors; Meta-analysis; Stroke

Introduction

Stroke is a serious global health problem, and considered as the fourth leading cause of death and the primary cause of adult disability in all around the world [1-3]. Sleep-related breathing disorders (SDB) is a frequent clinical phenomena, and characterized by repetitive apneas and hypopneas during sleep [4, 5].

The relationship of SDB with stroke has been investigated in the recent years. Many researches have reported increased incidence of SDB in stroke patients, which could lead to progression of the stroke [6, 7]. Furthermore, SDB following stroke may result in high mortality and morbidity [4]. Although SDB and stroke have common established risk factors including age, gender, elevated body-mass index (BMI), hypercholesterolemia, hypertension and habitual snoring [8-10], risk factors leading to SDB in stroke patients remains unclear. Thus, it is important to identify contributing risk factors or associated conditions that can impact on outcome of SDB following stroke.

Previously existing literature and primary analyses on risk factors of sleep disorder after stroke only involved in research of certain factor, and statistical power was deficient and insufficient in these studies [4, 11]. In this study, to investigate the risk factors for SDB after stroke and obtain a quantitative estimate of the risk, we systematically review the literature that have studied patients with stroke for SDB risk factors, (namely gender, Hypertension, Diabetes mellitus, Dyslipidemia, Smoking, alcohol use, Previous stroke history, Chronic obstructive pulmonary disease, Habitual snoring and Body-mass index)by conducting a meta-analysis.

Methods

Search strategy and study selection

We retrieved the relevant studies up to May 2014 in several public databases, mainly including PubMed, Medline, Springer, Elsevier Science Direct, Cochrane Library and Google scholar. The main search words were “factor” or “influence”, “affect” or “element” or “determinant” or “cause” or “reason” and “sleep disorder”, “apnoea” or “obstacle” or “impediment” or “barrier” or “obstruction” or “balk” and “stroke” or “cerebral apoplexy” or “acute cerebral accident” and “research” or “survey” or “study” or “trial”. Meanwhile, references from retrieved papers were checked for more relevant studies not identified in database search.

Eligibility of relevant studies

For relevant abstracts, full publications were retrieved for evaluation on the basis of criteria that were established a priori. All selected research articles were considered except case reports. We sought to include studies of any design that reported risk factors in sleep disorders, and relationships between sleep-disordered breathing, sleep apnea or insomnia symptoms and stroke without the limitation of sample size and range of age. Studies were excluded from the systematic reviews and reports that only described sleep disorder after stroke. Studies with no control group were also excluded. Reviews, reduplicated studies and studies published in language other than English were excluded as well.

Data extraction

Information from each study was extracted independently by two reviewers using a data extraction form. Study general characteristics (e.g., author, year of publication, location of participants, design of studies, ethnicity and number of cases), and characteristics of participants (e.g., age, gender and sample size) were recorded (where available) and doubled-checked. Meanwhile, the effect size of risk factors of sleep disorder in stroke patients compared with stroke patients without sleep disorder with corresponding 95% confidence interval (CI) were extracted as reported by authors. The effect size of gender, hypertension, diabetes mellitus, dyslipidemia, smoking, alcohol use, previous stroke, chronic obstructive pulmonary disease (COPD) and habitual snoring were measured by, and body-mass index (BMI) by weighted mean difference (WMD). In addition, we contacted authors of incorporated studies to obtain further information for data items that needed clarification. Discrepancies were resolved by discussing within our research team or contracting with the original investigators via e-mail.

Statistical analysis

The meta-analysis was focused on the risk factors of sleep disorder in stroke patients versus stroke patients without sleep disorder. The pooled estimates of effect size of risk factors were calculated by averaging OR or WMD based on a fixed or random effects model depending on the overall heterogeneity. Then, heterogeneity of effect size across studies was evaluated by applying Cochran’s Q-statistic and and the I2 statistic [12, 13]. P value < 0.10 or I2 value > 50% was considered to heterogenous across the studies. When substantial heterogeneity was detected, we calculated the overall estimate of ORs or WMDs using Mantel–Haenszel method in the fixed effect model [14]. If not, the pooled estimates were presented based on the random effect model by using DerSimonian and Laid method [15]. In addition, the publication bias were evaluated using Egger’s linear regression test [16], which measures funnel plot asymmetry by the natural logarithm scale of the effect size.

Meta-analysis was performed using the software of the STATA package v.11.0 (Stata Corporation, College Station, TX, USA). P value < 0.05 was considered to be statistically significant.

Results

Literature research and study characteristics

The details of the literature search were shown in a flow diagram (Figure 1). We obtained 972 citations (131 from Medline, 341 from Pubmed, 164 from Springer, 103 from Elsevier Science Direct, 16 from Cochrane Library and 217 from Google Scholar) with our electronic literature search. We obtain 84 citations after removing duplicates or irrelevant studies. In addition, 47 citations were excluded after screening based on the abstracts. Then 37 remained citations were full-text reviewed for detailed assessment. Finally, 8 citations satisfied the inclusion criteria and were eligible for the meta-analysis.

The characteristics and information of the included studies were presented in Table 1. The 8 selected studies contained 1381 patients with average age (ranging from 43.0 to 75.2 years old), including 578 experimental groups (stroke patients with sleep disorder) and 803 control groups(stroke patients without sleep disorder) [17-24].

Overall analysis of gender

As shown in Table 2, a total of 7 separate studies [17-20, 22-24], involving 1321 stroke patients (539 experimental groups and 782 control groups), were eligible for the meta-analysis. Meta-analysis showed no association between gender and sleep disorder after stroke was found (OR = 1.23, 95% CI = 0.72 to 2.1, P > 0.05), yet with

significant heterogeneity between studies (Q2 = 22.73, I2 = 73.6%, P < 0.1). In addition, the result of Egger’s test showed no evidence of publication bias (P = 0.91)

Overall analysis of hypertension

As presented in table 2, a total of 8 studies [17-24], including 1381 stroke patients (578 experimental groups and 803 control groups ), were finally eligible for the meta-analysis. Meta-analysis showed no association between hypertension and sleep disorder after stroke compared with control groups (OR = 1.22, 95% CI = 0.83 to 1.80, P > 0.05), yet with significant between-study heterogeneity was found (Q2 = 14.13, I2 = 50.5%, P < 0.1). In addition, the result of Egger’s test showed no evidence of publication bias (P = 0.43).

Overall analysis of diabetes mellitus

As shown in Table 2, a total of 7 studies [17, 19-24], involving 1326 stroke patients (549 experimental groups and 777 control groups), were eligible for the meta-analysis. Meta-analysis showed diabetes mellitus had a 41% increased risk of sleep disorder in stroke patients compared with control groups (OR = 1.41, 95% CI = 1.09 to 1.84, P < 0.05), yet with no significant between-study heterogeneity (Q2 = 1.59, I2 = 0%, P > 0.1). In addition, the result of Egger’s test showed no evidence of publication bias (P = 0.72)

Overall analysis of dyslipidemia

As shown in Table 2, a total of 4 studies [17, 19, 22, 24], involving 832 stroke patients (373 experimental groups and 459 control groups), were eligible for the meta-analysis. Meta-analysis showed no association between dyslipidemia and sleep disorder after stroke compared with control groups (OR = 0.96, 95% CI = 0.70 to 1.33, P > 0.05), yet with significant between-study heterogeneity (Q2 = 4.78, I2 = 37.3%, P > 0.1). In addition, the result of Egger’s test showed no evidence of publication bias (P = 0.68)

Overall analysis of smoking

As presented in Table 2, a total of 5 studies [17, 19, 20, 22, 23], containing 1234 stroke patients (495 experimental groups and 739 control groups), were eligible for the meta-analysis. Meta-analysis showed no association between smoking and sleep disorder after stroke compared with control groups (OR = 1.22, 95% CI = 0.83 to 1.80, P > 0.05), yet with significant between-study heterogeneity (Q2 = 13.82, I2 = 71.0%, P < 0.1). In addition, the result of Egger’s test showed no evidence of publication bias (P = 0.43).

Overall analysis of alcohol use

As shown in Table 2, a total of 5 studies [17, 19, 20, 22, 23], involving consisting of 1234 stroke patients (495 experimental groups and 739 control groups), were eligible for the meta-analysis. Meta-analysis showed alcohol use had a 41% increased risk of sleep disorder in stroke patients compared with control groups (OR = 1.59, 95% CI = 1.19 to 2.12, P < 0.05), yet with no significant between-study heterogeneity (Q2 = 4.49, I2 = 10.9%, P > 0.1). In addition, the result of Egger’s test showed no evidence of publication bias (P = 0.65).

Overall analysis of previous stroke

As shown in Table 2, a total of 5 studies [17, 19, 21, 22, 24], involving 892 stroke patients (412 experimental groups and 480 control groups), were eligible for the meta-analysis. Meta-analysis showed no association between previous stroke and sleep disorder after stroke compared with control groups (OR = 1.05, 95% CI = 0.74 to 1.49, P > 0.05), and with no significant between-study heterogeneity was found (Q2 =2.31, I2 =0%, P > 0.1). In addition, the result of Egger’s test showed no evidence of publication bias (P = 0.24).

Overall effects of chronic obstructive pulmonary disease (COPD)

As shown in Table 2, a total of 2 studies[19, 22], containing 623 stroke patients (280 experimental groups and 343 control groups), were eligible for the meta-analysis. Meta-analysis showed no association between COPD and sleep disorder after stroke compared with control groups (OR = 0.79, 95% CI = 0.40 to 1.57, P > 0.05), yet with no significant between-study heterogeneity was found (Q2 = 0.24, I2 = 0%, P > 0.1).

Overall effects of habitual snoring

As shown in Table 2, a total of 2 studies [20, 21], involving 422 stroke patients (138 experimental groups and 284 control groups), were eligible for the meta-analysis. Meta-analysis showed habitual snoring had a 1377% increased risk of sleep disorder in stroke patients compared with control groups (OR = 0.79, 95% CI = 0.40 to 1.57, P > 0.05), yet with no significant between-study heterogeneity was found (Q2 = 1.29, I2 = 22.6%, P > 0.1).

Overall effects of Body-mass index (BMI)

As shown in Table 2, a total of 3 studies [20, 21, 23], involving 494 stroke patients (176 experimental groups and 318 control groups), were eligible for the meta-analysis. Meta-analysis showed BMI had a 17% decreased risk of sleep disorder in stroke patients compared with control groups (WMD = 0.83, 95% CI = 0.63 to 1.03, P < 0.05), yet with no significant between-study heterogeneity was found (Q2 = 0.95, I2 = 0%, P > 0.1). In addition, the result of Egger’s test showed no evidence of publication bias (P = 0.998).

Discussion

In this study, we conducted a meta-analysis for investigating risk factors of sleep disorder in stroke patients. Based on the data from 8 studies including 1381 patients, the factors (Diabetes mellitus, alcohol use, habitual snoring and BMI) were associated with risk of sleep disorder in stroke patients when compared with control groups, yet with no significant between-study heterogeneity. The magnitude of this association was greatest for habitual snoring with stroke patients experiencing a 1377% increase in risk of sleep disorder compared with stroke patients without sleep disorder. However, there were no association between risk of sleep disorder after stroke and other factors (gender, Hypertension, Dyslipidemia, Smoking, Previous stroke history and COPD). In addition, there was no indication of publication bias from the result of Egger’s test.

Researchers have shown increased incidence of SDB in stroke patients, which may lead to a poor outcome and recurrence [25, 26]. SDB progresses from habitual snoring to the increased upper airway resistance syndrome to sleep apnea [27], and it is biologically plausible that habitual snoring could increase the risk of sleep disorder in stroke patients. Habitual snoring is significantly associated with daytime sleepiness, restless sleep, and hyperactivity, which could lead to sleep disorder [28, 29]. In our literature search we did not identify results of any randomized clinical trials evaluating the effect of risk factors on sleep disorder after stroke. However, it is reasonable to infer that habitual snoring may contribute to incidence and development of sleep disorder in stroke patients.

The previous studies have shown significant gender-related differences in presenting symptoms of patients with SDB [30], and the risk for the development of sleep disorder was significantly lower in women [31]. However, in our study, we found no association between gender and risk of sleep disorder after stroke, yet with the most significant between-study heterogeneities. This might result from unavoidable difference in study sample sizes and distinct adjustments across studies. Thus, we suggest that the present analyses might have underestimated the true effects of gender on risk of sleep disorder after stroke, and further studies are needed to be demonstrated this.

The present meta-analysis has a series of limitations, such as lack of uniformity (e.g. sleep disorder diagnostic criteria, the duration of patients with stroke) and exclusion of studies written in languages other than English. Furthermore, another limitation was the narrow coverage of information of the present analysis. After all, the results in this analysis were only obtained data from Asia, Europe, North America and Oceania, and the majority of the included studies did not state if there were any racial differences among participants. The differences in regions and races among studies might have an influence on pooled risk estimates of sleep disorder after stroke.

In summary, the present meta-analysis suggest that the factors (Diabetes mellitus, alcohol use, habitual snoring and BMI) is associated with risk of metabolic syndrome. The conclusion has a great significance for public health, especially in countries of high incidence of stroke. In addition, further investigations, both epidemiological and mechanistic, are needed to investigate whether regulating these factors can prevent occurrence and development of metabolic syndrome.

Order a Unique Copy of this Paper

Essay Creek is an academic writing service provided to you by, a London-based company.

  • Experience
    Helping students successfully for 11 years.
  • Confidentiality & Security
    Be sure your information will be kept confidential due to our secure service.
  • Quality & Reliability
    8.5 out of 10 average quality score according to our customers' feedback. 97.45% of orders delivered on time.
  • Versatility
    478 active writers in 68 disciplines.
  • 100% money back guarantee
    You can always request a refund if you are not satisfied with the result.

Read more about us

Our team of writers is comprised of people with necessary academic writing skills and experience in various fields of study.

  • Skilled writers only
    We carefully choose writers to employ, paying attention to their skills and abilities.
  • Competence
    Your order will be assigned to a competent writer who specializes in your field of study.
  • In-depth knowledge
    Our writers know both peculiarities of academic writing and paper formatting rules.
  • Motivation
    We keep updated on results our writers show, motivating them to constantly improve their performance.

Read more about our writers

  • Testimonials
    Our clients' testimonials prove we're doing everything right.

Check for yourself

  • Sample essays
    The best way to understand how well our writers do their work is to view sample essays written by them.

View samples

  • Our Free Essay Tools
    Even more opportunities to improve your academic papers.


Bibliography Generator
Words to Pages Converter
Words to Minutes Converter
College GPA Calculator
Thesis statement generator