The association between obstructive sleep apnea and lung nodule, carcinoembryonic antigen

The association between obstructive sleep apnea (OSA) and cancer risks gaining more and more attention. Data on the association between OSA and lung cancer risk are limited. This study is to investigate whether a link exists between low-dose computed tomography (LDCT) scanning of the chest findings, carcinoembryonic antigen (CEA) and OSA in patients suspected of OSA. The cross-sectional study included patients aged 18 years or older who underwent continuous nocturnal polysomnography at our sleep center between February 2019 and November 2020. All subjects underwent chest LDCT and CEA. Patients with an apnea–hypopnea index (AHI) of ≥ 15/h were classified as clinically significant OSA group, whereas patients with an AHI < 15/h were classified as control group. A total of 277 patients were enrolled in the study. 176 patients were categorized into the OSA group, while 101 patients were categorized into the control group. There is no relationship between any OSA-related parameter and presence of lung nodule or presence of ≥ 6 mm lung nodule in the binary logistic regression analysis. OSA group demonstrated a significant higher value of CEA than control group. Stepwise multiple linear regression analysis showed that lowest O2 saturation (β = − 0.256, p < 0.001), smoking status (β = 0.156, p = 0.007) and age (β = 0.153, p = 0.008) were independent predictors of elevated CEA. OSA was independently related to the elevated of serum CEA level, but not with presence of pulmonary nodule or ≥ 6 mm pulmonary nodule in LDCT. Further well-designed longitudinal studies with pathology available are needed to identify the association between OSA and risk of lung cancer.


Introduction
Obstructive sleep apnea (OSA), the most common form of sleep-disordered breathing (SDB), is mainly characterized by recurrent episodes of partial or complete upper airway obstructions during sleep that lead to intermittent hypoxia and sleep fragmentation. OSA is known to exert pathophysiologic and molecular disturbances on multiple organs and is linked to a variety of poor health outcomes. Recent evidences suggested that OSA was associated with increased incidence and mortality of cancer, especially in melanoma, lung cancer, prostate cancer, and colorectal cancer [1][2][3][4]. However, some studies opposed the results and indicated that OSA was not independently associated with cancer incidence and mortality [5]. It is not surprising that they commonly coexist in an individual as both OSA and cancer are common problems.

3
Lung cancer is the most commonly diagnosed cancer (11.6% of the total cases) and is the leading cause of cancer-related deaths (18.4% of the total cancer deaths) worldwide [6]. It is well known that tobacco smoking is a major risk factor for the development of lung cancer. There are several studies examining the association between OSA and the risk of lung cancer and its outcomes [7][8][9][10]. A study evaluating the prevalence of OSA in a lung cancer screening program found that OSA was very common in the high-risk population. Further multivariate analysis adjusted for confounders indicated that nocturnal hypoxemia was associated with positive screening findings (nodules > 6 mm) [11]. Another study investigating the prevalence of OSA in patients with newly diagnosed lung cancer showed a high prevalence of OSA (49%) in this population [12]. Nevertheless, the data of this issue are limited and the results remain inclusive. Furthermore, all studies focus on population with lung cancer or high risk of lung cancer. To our knowledge, there is no published study exploring this relationship based on population with OSA or clinical suspicion of OSA.
Low-dose computed tomography (LDCT) of the chest has been extensively used to screen lung cancer and shown to be efficacious at reducing mortality from lung cancer [13]. Serum tumor biomarkers can be accessed easily and economically, so they have been used as diagnostic adjunct and monitoring the course of lung cancer in the clinical setting. The National Academy of Clinical Biochemistry (NACB) recommends the serum CEA as a tumor biomarker for adjuvant lung cancer diagnosis. Hence, the present study aimed to explore the relationship between OSA and LDCT scanning of the chest finding in a cohort of patients studied for OSA suspicion. Furthermore, the association between OSA and CEA was also investigated in this study.

Subjects
Consecutive patients with age > 18 years undergoing overnight polysomnography because of symptoms of sleep apnea between February 2019 and November 2020 in our sleep center were included in the study. The exclusion criteria included those patients with known diagnosis of OSA and receiving any OSA-related therapies, patients with known diagnosis of any malignancy and receiving any tumor-related therapies, patients with serious comorbidities, such as acute myocardial infarction. Pregnancy was also excluded. The informed consents were signed by all parents. This study was reviewed and approved by ethics committee of the First Affiliated Hospital of Fujian Medical University (2020-422).

Anthropometric measurements and polysomnography
A detailed baseline evaluation including age, smoking habits, alcohol consumption, snoring, witnessed apneas, other medical history, and medication use were performed. Excessive daytime sleepiness was evaluated by the Epworth Sleepiness Scale. Body weight and height were measured in bare feet and light clothing in the morning by the same person using the same equipment. Body mass index (BMI) was calculated by dividing body weight by height squared (kg/m 2 ). Waist circumference (WC) was measured at the level midway between the lower costal margin and the iliac crest. Neck circumference (NC) was measured at the level of the laryngeal prominence. A full-night polysomnography (P Series Sleep System, Compumedics, Melbourne, Australia) was performed for every participant at the sleep center. All of the PSG data were visually scored by trained technicians according to standardized criteria [14]. Details have been described in our previous study [15]. The existing literature shows that there is a significant dose-response effect between AHI and the risk of cancer death; therefore, patients with an apnea-hypopnea index (AHI) higher than or equal to 15/h were categorized as having clinically significant OSA. Patients with an AHI less than 15/h were categorized into control group [16,17].

Blood samples
A fasting blood sample was drawn after a full night sleep study and then centrifuged at 3000 rpm for 10 min. CEA was quantitatively measured using electrochemiluminescence immunoassay (ECLIA) assay kits, according to the manufacturer's instructions (Cobas, Roche Diagnostics, Germany).

LDCT of the chest
All subjects were examined using a multilayer helical scanner at a low-dose project (105 mA, 120 kV, 2 mm overlapping sections). LDCT examination was performed in a single breath hold during complete inhalation. Chest CT image uploaded to Picture Archiving and Communication Systems (PACS). An experienced radiologist distinguished and described pulmonary nodules based on their size, presence, and type (solid, nonsolid, or partially solid) and the location of the nodules. It was reviewed by another senior radiologist. Pulmonary nodules was represented by the average the maximum nodule diameter and the nodule diameter perpendicular to this maximum.

Statistical analysis
The SPSS 19.0 package (SPSS Inc, Chicago, IL, USA) was used for the statistical analysis. GraphPad Prism 8 (Graph-Pad Software, USA) was used to draw the graphs. Normal distribution tests were performed on the data prior to analysis. In the case of normal distribution, the Student's t test was used to compare the means between two groups, or in the case of non-normal distribution, the Mann-Whitney U test was used for non-parametric data. Where appropriate, the categorical variables were compared using chi-square test or Fisher exact test. Binary logistic regression was used to analyze the relationship between OSA-related parameters and presence of pulmonary nodule, ≥ 6 mm pulmonary nodule. Correlations between variables were explored using the Pearson correlation test. Multiple linear regression was used to determine the independent predictors of CEA. When p value was less than 0.05, it was considered statistically significant.

Comparison of variables of all patients
The study group comprised 277 patients with 206 male and 71 female. The mean age was 53.49 ± 12.71 years, mean BMI was 27.34 ± 4.77 kg/m 2 , mean AHI was 28.89 ± 21.82. 101 patients with AHI < 15/h were categorized into control group, the remaining 176 patients with AHI ≥ 15/h were categorized into OSA group. Table 1 shows the comparison of baseline variables of OSA and controls. There were no significant differences in age, prevalence of diabetes, COPD, emphysema, smoking status, alcohol consumption between groups. The male sex ratio, prevalence of hypertension, dyslipidemia, BMI, NC, WC, AHI, oxygen desaturation index (ODI), the percentage of sleep time with SpO 2 (T90%), Epworth Sleepiness Scale (ESS) score of OSA group were significantly higher Out of 277 patients included, 143 patients with lung nodules were further analyzed. They were also divided into control group and OSA group. The detailed information is shown in Table 2. Similar to the results yielded from whole population, the prevalence of hypertension, dyslipidemia, BMI, NC, WC, sleep parameters and ESS score were significantly different between OSA and control group. There was no statistical difference in the prevalence of multiple pulmonary nodules, ground-glass opacity, nodule size or presence of lung nodule ≥ 6 mm between groups.

Relationship between OSA and lung nodules
Binary logistic regression was further performed to analyze the relationship between OSA-related parameters and presence of lung nodule, ≥ 6 mm lung nodule. The results are presented in Table 3. There is no relationship between any OSA-related parameter and presence of lung nodule or presence of ≥ 6 mm lung nodule. The correlation analysis indicated that the size of lung nodule was not correlated with any OSA-related parameter (Fig. 1). Figure 2 shows the comparison of CEA between OSA and control group. OSA group demonstrated a significant higher value of CEA than control group (2.53 ± 1.66 vs. 2.12 ± 1.15 ng/ml). The correlation analysis revealed that age, the presence of diabetes, tobacco status, BMI, WC, AHI, ODI, T90%, LaSO 2 , and

Discussion
In a cohort of patients with suspicion of OSA, we demonstrated that nocturnal hypoxia was independently related to the elevated of serum CEA level after adjusting confounding factors. OSA was not related with presence of pulmonary nodule or ≥ 6 mm pulmonary nodule in LDCT. There is growing interest in evaluating the association between OSA and cancer incidence. Several large cohort studies have investigated whether OSA increased the risk of cancer. A community-based cohort study with longterm follow-up found that moderate-severe OSA was significantly associated with incident cancer after adjusting confounding factors [18]. A multicenter, clinical cohort study including 4,910 patients investigated for suspected OSA revealed that severe OSA was associated with higher incidence of cancer, especially in patients aged < 65 years and men [2]. However, some study did not support an independent relationship between OSA and cancer incidence. Over a 7.8-year follow-up of 10,149 patients, Kendzerska et al. [1] found that the severity of sleep apnea was not significantly associated with either prevalent or incident cancer after adjustment for age, sex, BMI and smoking status at baseline. A recent prospective study with 8-year follow-up of 65,330 older women showed that OSA was associated with a 15% increase in total cancer risk. The association became non-significant after adjustment for multiple cancer risk factors [19].
Lung cancer is the most commonly seen cancer. Some studies focused on the relationship between OSA and risk of lung cancer and its mortality. A study performed in Germany reported high prevalence of OSA in newly diagnosed lung cancer patients, with the incidence of moderate to severe OSA reaching 17%. However, it could not determine whether OSA affects the prevalence of lung cancer [12]. Another study performed in Spain reported that the prevalence of OSA was 80% in consecutive patients with newly diagnosed lung cancer [9]. A cross-sectional study including 302 patients for investigating relationship between OSA and lung cancer discovered that AHI and nocturnal hypoxemia did increase the incidence of lung cancer, especially in moderate and severe OSA [7]. However, data are still limited in this issue and the conclusion is controversial.
Serum tumor markers are becoming more and more important in the diagnosis and staging of malignant tumors. CEA is one of the most commonly used tumor markers. Its overexpression is first found in colorectal cancer, and then in respiratory tract and other organs. Although the use of CEA in patients with lung cancer is still controversial, its low cost and easy availability make it a relatively reliable indicator of lung cancer. CEA is essentially a marker of adenocarcinoma, while it is closely related to TNM stage and tumor aggressiveness, so it has become an indicator of lung adenocarcinoma prognosis.
The mechanisms likely to be involved in the association between OSA and cancer included systemic inflammation, oxidative stress, angiogenesis, immune dysfunction. In an animal study, Almendros et al. [20] demonstrated that mice exposed to intermittent-hypoxia (IH) exhibited accelerated tumor growth and presented more tumor invasion. Tumorassociated macrophages from IH-exposed mice significantly increased TC1 migration and extravasation compared with normoxia group. Another animal study also confirmed that IH markedly enhanced the number and weight of lung metastatic colonies in mouse. Compared with control group, Lung tissues with tumor metastasis from IH group showed increased oxidative stress, as assessed by p22phox and  superoxide dismutase (SOD) mRNA levels and the NF-E2-related factor 2 (NRF2) protein Level, and inflammation, as assessed by Tumor Necrosis Factor-α (TNF-α) and interleukin-6 mRNA levels and the NF-κB P65 protein level [21]. Zhang et al. [22] showed that micro-vessel density (MVD) and vascular endothelial growth factor (VEGF) were significantly higher in tumor from IH animal than control group. Treatment with endostatin attenuated serum and tissue VEGF levels, lowering the MVD values. This study indicated that IH might accelerate tumor growth by facilitating the angiogenesis in tumor tissue. Our study revealed that OSA was not related with presence of pulmonary nodule or ≥ 6 mm pulmonary nodule in LDCT. In general, patients with ≥ 6 mm pulmonary nodule are considered to be at high risk of lung cancer, so we choose this standard. This is contradictory to the conclusion from a study, which indicated that nocturnal hypoxemia was associated with presence of > 6 mm lung nodules on multivariate analysis adjusted for confounders (OR = 2.6) [11]. The differences in the results might due to the different patient group. We included a group of patients suspected for OSA. The study by Perez-Warnisher et al. [11] included a group of patients with age and smoking criteria fulfilling a high risk of lung cancer (age 55−75 years; pack-years > 30). We also found that nocturnal hypoxia was independently related to the elevated of serum CEA level after adjusting confounding factors. This is consistent with a previous study reporting that the AHI score and ODI score were positively associated with the CEA level in T2DM patients [23]. Oxidative stress and inflammation responses are supposed to play a role in the association between OSA and elevated CEA [23,24]. The oxidative stress and inflammation responses induced by OSA trigger the transmission of oncogene signaling and lead to the development of cancer [21,25]. In addition, A previous study supported a positive correlation between serum CEA levels and lipid peroxidation [26]. Serum CEA levels are not only increased in lung cancer but also increased in other types of cancer, such as colorectal cancer. Actually, an previous epidemiological study showed that OSA was associated with an increased risk of colorectal cancer [27]. Therefore, it cannot be excluded that OSA was associated with other type of cancer (e.g., colorectal cancer).
However, it is vital to note the limitations of this study. First of all, this is a cross-sectional study; therefore, the observed associations do not indicate cause-effect relationships. Second, the sample size of this study is relatively small compared to the sample size of the community-based cohort study. Third, histological findings of lung nodules were not obtained in this study. Fourth, an AHI of 15/h, but not 5/h, was chosen as the threshold for grouping OSA and non-OSA. However, this threshold has been used in many previous studies and is also a useful limit to detect patients who need continuous positive airway pressure therapy. Finally, serum CEA level can be affected by many diseases or factors. Other cancer type such as colorectal cancer might not be excluded due to lack of endoscopic or PET-CT examining.
In conclusion, OSA was independently related to the elevated of serum CEA level, but not with presence of pulmonary nodule or ≥ 6 mm pulmonary nodule in LDCT. Further well-designed longitudinal studies with pathology available are needed to identify the association between OSA and risk of lung cancer.
Funding This research was funded by National Natural Science Foundation of China (No. 81900088 and 81870074) and Natural Science Foundation of Fujian Province (2022J011474). The sponsor had no role in the design or conduct of this research.

Data availability
The data sets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.