Key points are not available for this paper at this time.
Obstructive sleep apnea (OSA) is a complex and heterogeneous disease.1–4 More than one billion individuals worldwide are affected by OSA,5 which is defined as the number of apneas (nocturnal respiratory arrests lasting more than 10 s) plus the number of hypopneas (significant decreases in airflow accompanying oxygen desaturation or arousals of more than 10 s in duration) per hour of sleep.6,7 Sleep-disordered breathing results in both sleep fragmentation and continuous oxygenation–reoxygenation cycles (known as intermittent hypoxemia IH).8,9 The latter phenomenon (IH) triggers a series of pathophysiological mechanisms, primarily regulated by hypoxia inducible factor-1 (HIF-1),10 which increase the risk of cardiovascular events. These mechanisms include endothelial dysfunction,11 arteriosclerosis,12 reduced peripheral insulin sensitivity,13 a hypercoagulable state,14 and a higher incidence of arterial hypertension15,16 and diabetes.17 IH has also been linked to increases in neurocognitive,18 metabolic disorders19, and other comorbidities.20–23 In the last decade, a growing number of studies have explored the relationship between OSA and cancer prevalence, incidence, and aggressiveness. Early studies primarily analyzed this association with respect to the prevalence, incidence, and mortality of cancers as a collective group.24–31 Recently, systematic reviews and meta-analyses have analyzed these associations not only with all-cause cancers but also with specific tumor sites and histological types.32–42 Scientific evidence of this relationship has increased as a result of murine model studies43–45 and biomarker research,46–50 which have enabled the exploration of new pathophysiological pathways that could explain the results observed in clinical studies. This editorial has three objectives. First, it presents the main pathophysiological pathways described to date that support the hypothesis of a possible relationship between some types of cancer and OSA. Second, it analyzes recent systematic reviews and studies investigating the prevalence, incidence, and aggressiveness of cancer in general, and some types of cancer in particular, in relation to OSA. Finally, it places special emphasis on recently published studies that provide novel ideas regarding this relationship, especially with respect to the effect of OSA treatment with continuous positive airway pressure (CPAP). Future challenges are also discussed. All systematic reviews and meta-analyses indexed in PubMed (from 1950 to present) were systematically reviewed. The protocol was as follows: malig*Title OR tumorTitle OR cancer*Title OR melanomaTitle AND OSATitle OR apneaTitle OR CPAPtitle OR Positive airway pressureTitle restricted to meta-analyses and systematic reviews already published. Using these materials, a narrative review was conducted. The first pathophysiological mechanism postulated to explain the relationship between cancer and OSA is an increase in the expression of HIF-1 caused by IH secondary to respiratory events during sleep. This molecule is induced by hypoxia (both intermittent and sustained), and its main function is to increase oxygen delivery to hypoxic tissues. One of the main molecules whose expression and concentration increase in response to HIF-1 is vascular endothelial growth factor (VEGF), which promotes the neovascularization of hypoxic regions.50 While this neovascularization may be beneficial for certain cardiovascular events, it is hypothesized to have the opposite effect in cancer. Enhanced vascularization of tumor hypoxic regions in the context of IH due to OSA may accelerate tumor growth through increased cellular proliferation and promote hematogenous metastasis, ultimately contributing to greater tumor aggressiveness Supplementary Figure 1, https://links.lww.com/CM9/C368. This hypothesis has been corroborated by studies in animal models.43–45,51,52 IH has also been implicated in immune dysfunction through mechanisms such as the generation and propagation of reactive oxygen species (ROS), subsequent oxidative stress, and the onset of low-degree inflammation. These processes contribute to a cascade of events that progressively transform normal cells into malignant cancer cells. Among the immune system dysfunctions associated with IH, notable effects include macrophage polarization, natural killer T-cell deficiency, CD8+ T-cell lymphocyte dysfunction, CD3+ γσ-T-cell dysfunction, stem cell-like properties, and peripheral dendritic cell depletion.52–54 Recently, numerous studies have explored the role of IH in the promotion and function of some specific biomarkers related to cancer progression. In addition to HIF-1 itself, the most important of these biomarkers are those that induce neovascularization (such as VEGF) and other pro-inflammatory and pro-carcinogenic biomarkers. These include transforming growth factor (TGF)-beta 1, tumor necrosis factor (TNF)-alpha, tryptophan metabolism, cyclooxygenase-2 pathway, cannabinoid receptors, soluble programmed death-ligand 1 (sPD-L1), endostatin, endothelin-1 (and its receptors), oxidative stress molecules, and paraspeckle protein-1 upregulation.46–50 Ample evidence indicates that genetic factors influence the development of OSA, and an increased risk has been reported in first-degree relatives. Recently, well-powered genome-wide studies have suggested that the biological functions of genes involved in the OSA risk indicate the molecular mechanisms that also influence the risk of various types of cancer. The most consistently reported genetic findings connect OSA and cancer with variability in the HIF1A gene, which encodes the alpha subunit of transcription factor HIF-1.55,56 Tumor cells and neighboring cells within the tumor matrix communicate continuously through several processes, including paracrine and autocrine signaling, as well as through specific elements that mediate intercellular interactions. Recent studies suggest that exosomes are one of the key mechanisms underlying intercellular communication processes that govern alterations in the tumor microenvironment and directly affect the biological properties of cancer cells. Given their ubiquitous presence in bodily fluids and their ability to reliably deliver genetic messages between cells, ultimately altering the phenotype of the target cells, exosomes have emerged as potentially important biomarkers of end-organ morbidity in OSA. They also serve unique and important functions in solid tumors, potentially mediating the aggravated features of cancer in patients with OSA.57–60 In the last decade, there has been growing interest in the potential impact of microbial populations inhabiting the gastrointestinal tract on health and disease. Consequently, alterations in the composition and diversity of the microbiome (i.e., dysbiosis) have been implicated in a wide range of diseases, including cancer. In parallel with these advances in understanding the microbiome and cancer, similar associations have emerged linking changes in the gut microbiome induced by the presence of OSA to end-organ morbidities.61,62 The pathophysiological mechanisms described in clinical studies suggest an increase in the prevalence of cancer in patients with OSA. However, it is important to adjust for confounding variables that are risk factors for both diseases, such as age, smoking habits, and obesity. The most recent meta-analysis addressing this topic, published in 2022, comprised 22 studies involving 32.1 million individuals. It concluded that the adjusted prevalence of cancer in patients with OSA was 1.53 (95% confidence interval CI: 1.01–2.31) times higher than that of individuals without OSA.42 The first study to analyze the incidence of cancers with OSA was published in 2013 by Campos-Rodriguez et al,24 based on a database from the Spanish Sleep Network composed of 4910 individuals who attended sleep units for suspected OSA and were followed up for 4.5 years. This study concluded that, with a cutoff point of at least 43 events/h in the apnea-hypopnea index (AHI) (severe OSA), the incidence density ratio was 1.6 (95% CI: 1.2–2.1), compared to those with less than 18.7 events/h. Similarly, the incidence density ratio with nocturnal desaturation (more than 12% of the night time with an oxygen saturation below 90%) was 3.2 (95% CI: 2.3–4.6), compared to individuals with less than 1.2% of the night time with oxygen saturation below 90%. Subsequently, several authors have published results (either population-based or clinical) obtained from their patients’ cohort, albeit usually using a retrospective methodology. The latest meta-analysis on the risk of OSA in the incidence of all types of cancer was published in 2023; 18 studies were included, and it was observed that the incidence of cancer in patients with OSA (adjusted for several confounding variables) was RR: 1.36, 95% CI: 1.18–1.56, compared to patients without OSA. Subgroup analysis showed that the highest risk was among women (RR: 1.27, 95% CI: 1.06–1.51) and in the moderate-to-severe OSA groups (RR: 2.62, 95% CI: 1.64–4.19). Finally, considerable heterogeneity of results was observed, depending on the type of cancer studied, with a significant risk in breast cancer (RR: 1.32, 95% CI: 1.03–1.70); central nervous system cancer (RR: 1.71, 95% CI: 1.06–2.75); kidney cancer (RR: 1.81, 95% CI: 1.20–2.74); liver cancer (RR: 1.19, 95% CI: 1.10–1.29), and pancreatic cancer (RR: 1.23, 95% CI: 1.14–1.33) Supplementary Figure 2, https://links.lww.com/CM9/C368.32 The first analysis of mortality from all cancers and OSA was population-based and conducted in the Wisconsin cohort,25 consisting of 1522 patients followed up for 22 years. Compared with individuals without OSA, the adjusted relative hazards of cancer mortality were 1.1 (95% CI: 0.5–2.7) for mild OSA (AHI, 5.0–14.9 events/h), 2.0 (95% CI: 0.7–5.5) for moderate OSA (AHI, 15.0–29.9 events/h), and 4.8 (95% CI: 1.7–13.2) for severe OSA (AHI, >30 events/h). Therefore, there was a positive and significant linear relationship between OSA severity and cancer mortality. This study also replicated these results when analyzing the hypoxemia produced by respiratory events during sleep, suggesting that IH is the main cause of this increased risk of cancer incidence and aggressiveness in patients with OSA. Since then, other studies have analyzed cancer mortality, with contradictory results. However, the latest meta-analysis, which included 20 studies (5,340,965 participants), concluded that severe nocturnal hypoxemia nearly tripled all-cancer mortality (hazard ratio HR: 2.66, 95% CI: 1.21–5.85), although the AHI was not able to demonstrate any increase in mortality. These findings reinforce the role of hypoxemia in the relationship between OSA and cancer mortality.32 Cutaneous melanoma is possibly the cancer with the strongest association with OSA. In contrast to other cancers, nearly all published studies on the relationship between malignant cutaneous melanoma and OSA have provided positive results, and numerous animal and pathophysiological studies conducted specifically on this tumor have corroborated the clinical results. Although there is no clear explanation for this phenomenon, it is possible that malignant melanoma cells are more likely to proliferate in the presence of a higher concentration of HIF-1 generated by HI due to OSA. In the only prospective study conducted specifically in patients with melanoma in this context, 443 patients underwent a sleep study. Those with at least moderate OSA (AHI >15 events/h) presented with more aggressive melanoma (higher Breslow index) than those without OSA. This result was similar when oximetric variables were used in the assessment after adjusting for appropriate confounding variables.63 A recently published study based on the assessment of mortality or aggressiveness of certain types of cancer in the presence of OSA is worth highlighting. In total, 391 patients diagnosed with malignant melanoma were followed for 5 years. Of these, 137 did not present with OSA, whereas 52 deaths, 32 melanoma recurrences, and 53 cases of metastasis were reported during follow-up. After adjusting for age, gender, sentinel lymph nodes affected at diagnosis, body mass index, diabetes mellitus, mean saturation under 90% (TSat90), Breslow index, Epworth Sleepiness Scale, and melanoma treatment, moderate OSA (HR: 2.45; 95% CI: 1.09–5.49) and severe OSA (HR: 2.96; 95% CI: 1.36–6.42) were associated with a poorer prognosis of melanoma, compared with the control group. We will now delve into the relationship between melanoma and OSA as it is probably one of the most studied cancers in this context.64 Finally, a recent meta-analysis on this topic included six studies (both population-based and clinical) with more than 5.2 million individuals and concluded that the cumulative incidence was indeed higher in patients with OSA (HR: 1.71, 95% CI: 1.08–2.69).33 However, not all cancers have a clear association with OSA regarding incidence or aggressiveness. This is the case with lung cancer. One of the most discussed aspects is the great heterogeneity in the histological forms of lung cancer, and the different responses of each of these different cell lines to hypoxemia are primarily responsible for the discordance in the results. An interesting study conducted by Marhuenda et al65 induced normoxia, IH, or sustained hypoxemia “in vitro” in different lung adenocarcinoma and squamous lung cell carcinoma cell lines. They concluded that depending on the etiology (lung adenocarcinoma or squamous lung tumor), the behavior after exposure to IH (typical of OSA) and sustained hypoxemia (typical of chronic cardiopulmonary diseases such as heart failure or chronic obstructive pulmonary disease) was different in the samples exposed to normoxia. Even within different lung adenocarcinoma cell lines, there were differences in response to IH and HC, with the H522 cell line presenting greater expression of epithelial cell adhesion molecule (EpCAM), which has been associated with lung cancer and survival. A recent meta-analysis on the association between the incidence of lung cancer and OSA has pointed out this heterogeneity in the studies. However, after including seven studies with almost 5 million individuals, mainly derived from epidemiological series, it is concluded that OSA was associated with a higher incidence of lung cancer (HR: 1.25; 95% CI: 1.02–1.53).36–38 Breast cancer is the most common cancer in women. Although the characteristics of OSA in women differ from those in men, there is significant histological and immunohistochemical variability in breast tumors, similar to that observed in lung cancer.63 A recent meta-analysis that included six studies with more than 5 million individuals concluded that patients with OSA were 36% more likely to develop breast cancer during their follow-up (HR: 1.36; 95% CI: 1.03–1.80) compared to those without OSA, after adjustment for age, gender, obesity, diabetes mellitus, alcohol consumption, and hypertension.35 However, one study conducted in patients with breast carcinoma found no relationship between OSA and any of the four breast cancer types studied.66 Finally, no association with OSA has been demonstrated in other common cancers, such as colorectal cancer. More specifically, a 2023 meta-analysis of this association, which included four studies with more than 5.6 million individuals, did not find a higher incidence of colorectal cancer in patients with OSA than in those without OSA (HR: 1.49; 95% CI: 0.75–2.97).39 Recent epidemiological or prospective studies on some tumor types have analyzed the possible effect of CPAP treatment (the treatment of choice in patients with symptomatic or moderate-to-severe OSA) on tumor incidence or aggressiveness after adjusting for several confounding variables. In the aforementioned study of 391 melanoma patients followed for 5 years, it was observed that patients with OSA who were treated with CPAP and showed good adherence to treatment (at least 4 h per night) significantly reduced their risk of poor melanoma prognosis to levels similar to those of patients without OSA.64 However, a study based on a French health administration database, including 4499 individuals followed for a median of 5.4 years, observed that 194 developed cancer, with no significant difference between individuals with OSA who were prescribed CPAP and demonstrated good adherence to this treatment and those who were not prescribed CPAP or did not demonstrate good adherence.67 The worldwide prevalence of both cancer and OSA is extremely high. Therefore, investigating an association between these two disorders is of enormous scientific interest, especially when the presence of OSA could be a risk factor for some cancers (especially as a consequence of IH) and a potentially treatable trait in some cancers. However, many questions about this topic remain, including: Why is there so much heterogeneity in the results, primarily depending on the cell line and location of the tumor? What is the potential role of CPAP in eliminating respiratory events during sleep in patients with OSA? How should the different confounding variables that are obvious risk factors for both diseases (e.g., age, sex, smoking, and obesity) be managed?
Oscullo et al. (Tue,) studied this question.