Obstructive sleep apnea (OSA) is a sleep-related breathing disorder characterized by obstruction of the upper airway during sleep, leading to fragmentation of sleep and intermittent hypoxia with episodes of apnea or hypopnea and increased respiratory effort.¹ OSA is associated with several comorbidities, including cardiovascular,² neurological,^3-5 and endocrinological⁶ disorders. Metabolic syndrome, characterized by central obesity, decreased plasma high-density lipoprotein, elevated blood pressure, and increased fasting plasma glucose and plasma triglyceride (TG),^7-9 is more common and severe in patients with OSA.^10,11 The appropriate and effective treatment of OSA using positive airway pressure (PAP) therapy results in improvement of many metabolic syndrome parameters, confirming the strong association between OSA and metabolic syndrome.^12-15 Atik et al.¹⁴ showed that 3 months of PAP treatment significantly decreased total cholesterol, low-density lipoprotein (LDL) cholesterol, and glycosylated hemoglobin (HbA1c) levels in OSA patients. In agreement with these findings, Dorkova et al.¹⁵ showed that PAP therapy reduced total cholesterol and homeostatic model assessment of insulin resistance (HOMA-IR) in addition to systolic and diastolic blood pressure. On the other hand, limitations related to control of pre-existing comorbid diseases or concomitant risk factors are debatable in the cohorts of these studies.¹⁶

Smoking is an important relevant concomitant risk factor for both OSA and metabolic syndrome. Smoking increases the risk of metabolic syndrome and related systemic complications up to 2.5 to 3-fold.¹⁷ Smoking also exacerbates OSA by causing chronic inflammation and neuromuscular dysfunction in the upper airways¹⁸ and decreases PAP adherence possibly because of the more frequent side effects associated with the upper airway.¹⁹ To the best of our knowledge, the effect of active smoking on the baseline parameters of metabolic syndrome after PAP treatment in patients with OSA has not yet been reported. In the present study, we hypothesized that active smokers with OSA would benefit less from PAP treatment compared with non-smoker patients with OSA in terms of favorable metabolic changes. Therefore, the effects of active smoking on the parameters of metabolic syndrome in OSA patients treated with PAP therapy were investigated.

Patient selection

This was a prospectively designed study to analyze the clinical, laboratory, and polysomnography (PSG) data of patients admitted to our Sleep and Disorders Unit with symptoms/signs of sleep-disordered breathing. One-night PSG was recorded and scored according to the latest version of the “Manual for scoring sleep and related events” by the American Academy of Sleep Medicine,²⁰ and the diagnosis of OSA was based on the latest version of the International Classification of Sleep Disorders (ICSD-3).²¹ All patients >18 years of age with an apnea-hypopnea index (AHI) ≥15 per hour, who agreed to participate and provided written informed consent, were consecutively included in the study. Exclusion criteria were previous or current use of PAP therapy; pregnancy; malignancy, type 1 diabetes mellitus, chronic obstructive pulmonary disease, cardiovascular disease, dementia, or psychiatric disorder; insulin therapy, anti-diabetic, anti-hypertensive, or anti-hyperlipidemic treatment; and alcohol and/or substance use/abuse. The local ethics committee approved the study (approval number: 83045809/8252). A detailed informed consent form was obtained from all study participants.

Clinical evaluation

All patients meeting inclusion and exclusion criteria were clinically examined, and the smoking status of the patients was recorded as active smoker or non-smoker. Active smoking was defined as inhaling mainstream smoke, including occasional use in social events. Patients who had never smoked were accepted as non-smokers. Subjects exposed to secondhand smoking, defined as passive environmental tobacco smoke exposure and individuals who had quit smoking, were excluded. Electronic cigarette users were also excluded from the study because the effects and dosages could not be reliably quantified. Demographic and anthropometric data of the patients including age, gender, body mass index (BMI), as well as neck, waist, and hip circumferences were also recorded.

Biochemical tests

Fasting blood samples were obtained from all participants in the morning following PSG recording, and the following parameters were studied:

(1) Plasma glucose (using the hexokinase method)

(2) HbA1c (using high-pressure liquid chromatography)

(3) Serum insulin (using the electrochemiluminescence immunoassay [ECLIA] method)

(4) Oral glucose tolerance test (OGTT) to measure plasma glucose levels at the first and second hours following 75 g of oral glucose administered after 10 hours of fasting

(5) Plasma TG and LDL (using an enzymatic method in an automatic analyzer)

(6) Serum leptin (using the enzyme-linked immunosorbent assay [ELISA] method)

(7) HOMA-IR index [fasting plasma glucose (mmol/L)×fasting serum insulin level (mIU/L)/22].

Follow-up evaluations

All participants were enrolled for manual PAP titration in the sleep laboratory, and the therapeutic measures were adjusted using the recent international guidelines.²² All patients were strictly advised to comply to PAP therapy, and a follow-up examination was planned for three months after initiation of therapy. All patients were advised to continue their habitual living conditions without major changes during the study period.

At the follow-up examination, compliance with therapy was evaluated based on the output of the PAP machine. Patients who used PAP therapy effectively (for at least five nights in a week and for at least 4 hours per night corresponding to approximately 70% of nights per week) were accepted as compliant participants based on the definition by the Centers for Medicare and Medicaid Services.¹⁹ Patients who did not continue with PAP therapy for 3 months or subjects who used PAP therapy ineffectively were non-compliant and were excluded from the study (Fig. 1). Clinical examination, anthropometric measures, and biochemical tests were all repeated for the compliant participants. A change in parameter was accepted as stable when it remained within the limits of ±1 standard deviation (SD) and as increased or decreased when the change was more than ±1 SD.

Statistical analysis

Data were analyzed using SPSS version 15.0 (SPSS Inc.) software. Nominal parameters were compared using the chi-square test, and non-parametric continuous parameters were compared using the Mann-Whitney U-test. Subgroup analyses to compare pre- and post-treatment values were performed using the Wilcoxon test (due to non-parametric distribution of the parameters). The Mantel-Haenszel test was performed to adjust the confounders in the chi-square test. Adjustment for multiple comparisons in the Mann-Whitney U-test were measured using one-way analysis of variance (ANOVA) and post hoc least significant difference tests. Spearman correlation test was performed in the analysis of relationships between biochemical parameters. Patients with OSA were subgrouped as obese (BMI >30 kg/m²) or non-obese (BMI <30 kg/m²) and subjected to Spearman correlation analysis to evaluate the relationship among parameters of metabolic syndrome in both active smokers and non-smokers. A P<0.05 was considered statistically significant.

The present study included 115 patients (72 males, 62.6%); 53.3% (n=61) were active smokers, and the smoking percentage was similar between male and female patients with OSA (52.4% vs. 55.6%, P=0.596). The mean age of the study population was 53.4±14.3 years, and active smokers were significantly younger than non-smokers (P=0.034) (Table 1). The mean BMI was 31.5±4.6 kg/m², mean neck circumference 41.4±4.1 cm, mean waist circumference 110±13.4 cm, and mean hip circumference 112.6±8.4 cm. Anthropometric measurements did not significantly differ between active smokers and non-smokers (Table 1).

In PSG recordings, the mean AHI was lower in active smokers than in non-smokers; however, the difference was not significant (32.8±20.1/hour vs. 47.2±25.6/hour, P=0.169). The ratio of severe OSA in non-smokers was significantly higher than in active smokers (P=0.049) (Table 1). All other PSG parameters were similar between the active smoker and non-smoker patients (data not shown).

At the follow-up examination after three months of PAP therapy, neither clinical nor anthropometric parameters showed significant changes (data not shown). In biochemical parameters, changes in plasma TG and LDL levels, fasting plasma glucose, OGTT at hour 1 and 2, HOMA-IR index, blood HbA1c level, and serum insulin and leptin levels did not show significant difference in active smoker and non-smoker OSA patients (Table 2).

Conversely, in analysis of stable, decreased, or increased parameters (Table 3), serum leptin level was decreased in 78.6% of non-smoker OSA patients after 3 months of treatment and decreased in only 46.7% of active smoker OSA patients despite effective treatment (P=0.029). This significant difference was confirmed after adjusting for confounding factors, including gender, age, and weight, using the Mantel-Haenszel test (P=0.034). Furthermore, leptin level was increased in 40% of active smoker OSA patients but not in any of the non-smoker OSA patients. Leptin levels before and after PAP therapy in active smoker and non-smoker patients with OSA are shown in Fig. 2. Spearman correlation analyses showed that elevated leptin level in active smokers positively correlated with TG and HOMA-IR, both in pre- (P=0.374 and P=0.367, respectively) and post-treatment (P=0.252 and P=0.232, respectively) measurements, although they were not significantly different. In subgroup analyses, leptin level in obese (BMI ≥30 kg/m²) active smoker OSA patients positively correlated with HOMA-IR at follow-up examination after PAP treatment but also without significance (P=0.062).

The main and original finding of the present study was that active smoking in OSA results in elevated leptin level despite effective PAP therapy, even after adjusting for important confounders. OSA is associated with increased serum leptin level,^23,24 and treatment of OSA with PAP therapy results in reduction of serum leptin level.^25,26 In a meta-analysis of 11 studies with 413 participants by Chen et al.²⁷, PAP treatment significantly reduce serum leptin level without concomitant weight loss, starting from the 3rd day of the treatment and lasting for 3 months and beyond. Similarly, in the present study, a significant decrease in leptin level was observed in non-smoker OSA patients after months of effective PAP therapy, without any significant changes in body weight or BMI. However, in active smokers, the effect of PAP treatment was blunted, and leptin level further increased.

Smoking impairs insulin signaling via intramuscular lipid accumulation due to lipodystrophy in adipose tissues,²⁸ which explains the lack of benefit from insulin-sensitizing effects following weight loss in smokers. Smoking was suggested as a predictor of increased serum leptin level independent of gender, BMI, or AHI in OSA patients.²⁹ Although PAP treatment was shown to effectively reduce serum leptin level,^25-27,30 contradictory data were also reported in the literature. In one study, serum leptin and adiponectin levels were not significantly changed in OSA patients with coronary artery disease after PAP treatment.³¹ Furthermore, an increase in plasma leptin level in parallel with an increase in waist circumference was observed after 12 months regardless of PAP treatment, and lifestyle modifications were suggested. Therefore, confounding factors, such as smoking, should be considered when evaluating the effects of PAP treatment on metabolic parameters.

Leptin plays an important role in energy balance and lipid storage, and increased levels are associated with obesity and insulin resistance.^32,33 In the present study, HOMA-IR was increased as leptin level increased; although this positive correlation was not significant, it was more pronounced in obese patients. Conversely, a positive and significant correlation between leptin and insulin resistance (HOMA-IR) was observed after adjusting for age, sex, smoking, BMI, and TG level.³² In addition to obesity, direct effects of leptin on the hypothalamus were proposed as an underlying pathophysiologic mechanism of metabolic consequences.³⁴ Elevated serum leptin level was shown to play an important role in many hormonal, metabolic, thrombotic, and inflammatory processes, leading to further systemic complications.^35-38 Therefore, elevated leptin level could potentially lead to secondary adverse effects during OSA, which is also an important contributor to increased leptin level. This bidirectional relationship between OSA and leptin emphasizes the importance of an effective OSA treatment using PAP therapy as well as rational therapeutic approaches including lifestyle changes and habitual modifications such as body weight loss and smoking cessation.

The main limitations of our study were the small number of participants and short follow-up duration. Long-term data from a larger cohort would provide additional information regarding the multifactorial and complex relationship between OSA and leptin in terms of confounding factors such as smoking and the effects of PAP treatment. In addition, a dose-response relationship between smoking and OSA was reported in the literature,³⁹ and total smoking exposure and smoking intensity should be measured as important confounding factors in future studies. Furthermore, because leptin is a hormone secreted by adipose tissue, adjusting leptin level based on body fat mass would provide a better estimation of the relationship between smoking and metabolic syndrome components before and after PAP therapy. The lack of data on body fat mass as well as dietary intake that may also affect serum leptin level is another important limitation of the present study. These emerging questions require further research for more comprehensive understanding.

In conclusion, the results of the present study showed that serum leptin level was stable or decreased in non-smokers with OSA after PAP therapy and increased in 40% of active smokers. Because high serum leptin level is closely associated with leptin resistance characterized by increased central leptin expression and decreased leptin receptors, smoking may play a predisposing role in leptin resistance in patients with OSA despite PAP therapy. To better understand leptin resistance in smoker and non-smoker subjects with OSA, further biomarker studies are needed on the metabolic consequences of leptin resistance in active smoker and non-smoker patients with OSA.

The authors declare no conflict of interest.

This study was funded by Istanbul University Scientific Research Projects Department (Project number: 31684).

Study concept and design: GBS; acquisition of data: MAS, VK, and GBS; analysis and interpretation of data: MAS, VK, and DK; drafting of the manuscript: MAS; critical revision of the manuscript: GBS and DK; statistical analysis: MAS and VK; obtained funding: VK and DK; administrative, technical, or material support: MAS, VK, and GBS; and study supervision: DK.