Nonalcoholic fatty pancreatic disease and cardio-metabolic risk: is there is a place for obstructive sleep apnea?

Recent data strongly suggest that OSA is an important contributing factor for the occurrence of dyslipidemia [29, 30]. The pathogenesis of the above association is mediated by at least three major processes such as enhanced lipolysis in adipose tissue, decreased lipid clearance and upregulated lipid synthesis in the liver. A review of the literature on NAFLD and OSA by Mirrakhimov and Polotsky noted factors that play key pathobiological roles in the pathogenesis of NAFLD in the setting of OSA [5]. Intermittent hypoxia (IH) is a key mediator of OSA and leads to various alterations in biological homeostasis [31]. First of all, it activates the sympathetic nervous system via the carotid body chemoreceptors, which in turn leads to fat degradation (lipolysis) from adipose tissue. Free fatty acids (FFA) compose the triglycerides (a major form of fat storage in the body; another major constituent of triglycerides is glycerol), which are transported to different organs to serve as fuel. However, FFA can lead to steatosis and organ inflammation, such as in the setting of NAFLD. At the same time, IH activates hypoxia inducible factors 1 and 2 (HIF-1 and HIF-2 respectively), which lead to decreased lipid metabolism in the liver, increased hepatic fat synthesis, upregulated liver inflammation and fibrosis. The HIF family is thought to play a major role both in physiologic adaptation as well as in the pathophysiology of disease states [32]. Indeed, the activation of HIF seems to play a key role in the development and pathogenesis of cancer [32‐35]. Therefore, the previously described relationship between OSA and cancer [8‐11] may be mediated via the HIF pathway.

More recently a study by Minville et al. showed that OSA and more specifically, nocturnal hypoxia was demonstrated to be independently associated with liver steatosis in obese individiuals further supporting the independent interrelationship between OSA and metabolic diseases [19]. Furthermore, a recent meta-analysis by Sookoian and Pirola showed that patients with OSA have greater transaminase values, higher prevalence of NAFLD and liver fibrosis compared to individuals without OSA [20].

An important question is whether OSA is independently (adjusted for confounders such as body mass index (BMI), smoking, age, gender, alcohol intake, family history etc.) associated with organ-specific cancers [11]. Of particular importance is to study whether OSA is linked to hepatocellular carcinoma via its association with the occurrence of NAFLD.

We hypothesize that OSA may contribute to the occurrence of NAFPD in a similar fashion as it does to NAFLD. Future studies should assess whether OSA is a true independent risk factor for the occurrence of NAFPD, given that NAFPD may precede NAFLD [20] and thus, may be an early marker of excess in cardiac and metabolic risk. Therefore, identifying patients with NAFPD and OSA even without clinically overt metabolic disease such as metabolic syndrome or type 2 DM would necessitate aggressive risk reduction in order to prevent the future development of type 2 diabetes mellitus, which is a major contributor to excessive morbidity and mortality worldwide.

The relationship between metabolic health and fatty infiltration of the pancreas is a well-known phenomenon, which was first described in 1926 by Schaefer, who showed a correlation between the weight of the human body and that of the pancreas [36]. In 1933 Ogilvie demonstrated that obese cadavers had heavier pancreatic glands than non-obese cadavers [37]. Both of these pioneer studies suggested that pancreatic weight goes in parallel with body weight. It seems that this association may now be explained by ectopic fat deposition in pancreatic tissue.

Several animal studies have investigated the effects of fat deposition in pancreas. Shimabukuro et al. demonstrated that FFA induced apoptosis of pancreatic beta- cells (endocrine cells producing insulin) in Zucker diabetic fatty rats [38]. Mathur et al. showed that obese mice had heavier pancreata with excessive fat infiltration and enhanced production of pro inflammatory cytokines [39]. Lee et al. demonstrated fat accumulation in both the exocrine and endocrine pancreas in a rat model [40]. These researchers speculated that fatty infiltration of the endocrine pancreas may impair insulin secretion and thus lead to type 2 DM.

Tushuizen et al. studied 12 males with type 2 DM and 24 age- and BMI-matched non-diabetic men [41]. These researchers demonstrated that pancreatic fat was independently (adjusted for BMI, fasting plasma glucose and triglycerides) associated with worse pancreatic beta cell function. Lee et al. showed that NAFPD was associated with higher levels of insulin resistance, visceral adiposity, triglycerides and alanine aminotransferase compared to normal pancreata [22].

Subsequently, scientific studies showed that increased age, obesity, visceral steatosis, hypertension and type 2 DM were associated with the presence of NAFPD [42‐45]. Nevertheless, it is essential to admit that NAFPD is understudied entitiy and not much is known about the natural course of NAFPD. Available evidence suggests that men have a higher prevalence of NAFPD compared to women with the same BMI, insulin resistance and fat intake [46, 47]. Certain groups may also have a greater prevalence of NAFPD. For example it was shown that Hispanics had greater pancreatic fat content than African-Americans [48]. These observations may explain why men and certain ethnic groups have increased metabolic risk profiles.

Most studies to date have investigated patients at risk for type 2 DM and, thus, the prevalence may be overestimated. Several imaging modalities can be of use in quantifying pancreatic fat such as trans abdominal and trans esophageal ultrasound, computed tomography and magnetic resonance tomography [44‐46, 49‐51]. However, NAFPD is currently the focus of research rather than the clinical community given scarce data on natural history and information on whether screening for NAFPD will translate into decreased morbidity and mortality.

In a pig model Hanukkainen et al. demonstrated that pancreatic fat is associated with insulin resistance and liver fat content [52]. However, van Geenen et al. showed that the association between pancreatic fat and liver fat is mediated via central obesity [53]. Heni et al. demonstrated that pancreatic fat was negatively associated with insulin secretion in patients with pre-diabetes [54]. These researchers suggested that this effect may contribute to the progression of pre-diabetes to overt type 2 DM. van der Zijl et al. also demonstrated that patients with pre-diabetes had greater pancreatic fat content [55]. However, this group was unable to demonstrate a relationship between pancreatic fat and insulin secretion. On the other hand, insulin resistance was found to be strongly associated to the presence of pancreatic fat [56]. Sepe et al. found a close interrelationship between NAFPD and metabolic syndrome [24]. However, no correlation was found between NAFPD and pancreatitis, pancreatic cancer and DM.

More recently two studies from Taiwan were published on the association between NAFPD and metabolic risk. In a first study, Wu and Wang confirmed an intricate association between NAFPD and metabolic risk factors [21]. They retrospectively studied 557 subjects without known hypertension and DM for the prevalence of NAFPD. 12.9 % of subjects were found to have NAFPD via transabdominal ultrasonography. Older age, greater BMI, abdominal waste circumference and metabolic variables such as glycated hemoglobin, lipid parameters and systolic blood pressure were associated with the presence of NAFPD. In a second study, Ou et al. enrolled 7,464 individuals to assess prevalence of NAFLD and NAFPD [57]. These researchers found that NAFPD was associated with pre-diabetes and DM after adjustment for cardiometabolic variables. It is interesting to note that NAFPD was only found to be associated with the presence of pre-diabetes and DM in male subjects, but not in females. The lack of association between female gender, NAFPD and metabolic risk may be explained by the study methodology, lower prevalence of NAFPD compared to NAFLD and possible gender specific patterns of ectopic fat deposition. A summary of key clinical studies investigating fatty pancreas is presented in Table 1.

Future research is needed to clarify whether NAFPD is a simple consequence of obesity or indeed, an independent player in the development of metabolic complications such as type 2 DM. It will also be interesting to investigate whether the presence of NAFPD is related to the severity of clinical pancreatitis. Of greater importance is whether NAFPD represents a risk factor for the occurrence of exocrine pancreatic cancer.

We postulate that NAFPD is an early marker of increased cardiac and metabolic risk. NAFPD may precede clinically overt metabolic disease (such as NAFLD, metabolic syndrome and type 2 DM) and, thus, would be a useful marker of cardiometabolic risk in any individual. Therefore, identifying such individuals is of paramount importance since early intervention may reduce the burden of type 2 DM and cardiovascular disease. On the other hand, OSA was shown to be a risk factor for a myriad of cardiac and metabolic complications such as cardiovascular disease, hyperlipidemia, type 2 DM and NAFLD. Of particular note, NAFLD is a spectrum of metabolic syndrome and is related to a greater risk of vascular disorders. It was reported that OSA is a contributing factor for the development of NAFLD, and through this mechanism may lead to a greater risk of type 2 DM. Furthermore, OSA may exert hypoxic damage to pancreatic β-cells and lead to the occurrence of type 2 DM in susceptible individuals.

Further studies are needed to clarify whether OSA is related to NAFPD. Studies assessing the prevalence of NAFPD in patients with OSA (and vice versa) may be needed initially. The results of such studies should be adjusted for common confounders such as age, gender, BMI and obesity, metabolic syndrome, type 2 DM, hyperlipidemia and alcohol intake. Subsequently, longitudinal studies of OSA patients will be needed to assess the impact of OSA on pancreatic fatty infiltration and the natural history of NAFPD. Studying the natural history of NAFPD (with and without concomitant OSA) is of paramount importance and such research should give answers to the following question: is NAFPD truly a benign disease or is it associated with subclinical pancreatic inflammation and a resultant increase in the severity of clinical cases of acute pancreatitis, pancreatic insufficiency (both exocrine and endocrine) and pancreatic cancer. In turn, it should be determined whether OSA is a key player in the development of the above complications.

These research questions somewhat parallel those which have been asked about NAFLD, a sister disorder of NAFPD. NAFLD was found to be a non-benign disease, with some patients progressing to liver cirrhosis and hepatocellular cancer. Whether the same is true for NAFPD is a matter for future research.