Obstructive sleep apnea (OSA) has been suggested as an independent risk factor for nonalcoholic fatty liver disease (NAFLD), and continuous positive airway pressure (CPAP) is the first-line therapy for OSA.

To clarify the efficacy of effective CPAP therapy on NAFLD of OSA patients by serum markers and transient elastography (TE) using FibroScan^® (Echosens, Paris, France).

We prospectively enrolled 123 consecutive patients with OSA who met the indications for CPAP. Liver fibrosis and steatosis were assessed using TE. Before and after 6 mo of CPAP therapy, serum markers and TE were assessed for all patients. The mean usage rate of CPAP therapy for 6 mo was arbitrarily calculated in each patient and expressed as “mean compliance index” (m-CI).

In 50 OSA patients with NAFLD, both aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels were significantly decreased after 6 mo of CPAP therapy. Univariate analysis showed that decreased body weight (BW), decreased body mass index (BMI), decreased AST level, decreased hemoglobin A1c, and high m-CI were significantly related with improved ALT level. In multivariate regression model adjusted for quantities of BW change during 6 mo of CPAP therapy, high m-CI tended to improve ALT level (P = 0.051). All 17 OSA patients with NAFLD, high m-CI and no BMI changes showed significant improvements in AST and ALT levels. Meanwhile, no significant changes in TE data or serum fibrosis markers were seen.

Some NAFLD could be associated with chronic intermittent hypoxia due to OSA independent of BW changes. In those cases, adequate reoxygenation from effective CPAP therapy may improve NAFLD.

Nonalcoholic fatty liver disease (NAFLD) seems to be becoming the most common liver disease worldwide as virus-related liver diseases are being successfully addressed. NAFLD affects 20%-40% of the general population in Western countries[1,2], and 9%-30% of the Japanese population[3,4]. Overnutrition and/or lack of exercise leads to metabolic syndrome, including visceral obesity, impaired glucose tolerance, dyslipidemia, and hypertension. NAFLD is also considered as the hepatic manifestation of metabolic syndrome. However, whether NAFLD represents a risk factor for cardiovascular disease and cerebral vessel disease remains unclear[5-8]. NAFLD encompasses a large spectrum of histological liver alterations, ranging from simple steatosis to steatosis accompanied by hepatocellular injury, inflammation, and fibrosis, termed nonalcoholic steatohepatitis (NASH). NASH has the potential to progress to cirrhosis, which is an established risk factor for hepatocellular carcinoma. In this setting, NAFLD is considered a major public issue all over the world, and improved treatment is imperative. However, the pathogenesis is still poorly understood, and effective pharmacotherapy as an adjunct to lifestyle modifications has yet to be established.

Obesity is the highest risk factor for NAFLD and is known to be frequently complicated by obstructive sleep apnea (OSA). OSA is a common medical condition characterized by repetitive partial or complete obstruction of the upper airway, resulting in increased negative intrathoracic pressure, sleep fragmentation, and chronic intermittent hypoxia (CIH) during sleep. Several experimental studies have demonstrated that CIH from OSA may be relevant to the presence and severity of NAFLD[9-11]. Moreover, OSA is independently associated with an increased prevalence of metabolic syndrome[12] and also leads to chronic systemic inflammations, insulin resistance, and liver injury. Repetitive hypoxia/reoxygenation during the transient cessation of breathing in OSA resembles ischemia/reperfusion injury and is known to promote production of reactive oxygen species, leading to oxidative stress[13]. Savransky et al[10] demonstrated that CIH in lean mice led to lipid peroxidation of liver tissue and increased levels of the active proinflammatory transcription factor nuclear factor kappa βin the nuclear fraction of hepatocytes, suggesting that CIH induces oxidative stress in the liver[10].

Continuous positive airway pressure (CPAP) therapy can treat upper airway collapse during sleep, and is now accepted as the standard and first-line therapy in the current management of OSA. The benefits of CPAP therapy have been demonstrated and established in terms of decreasing hypoxic events and daytime sleepiness, and of ameliorating metabolic and cardiovascular diseases[14-16]. Whether CPAP therapy improves the liver injury of NAFLD with OSA has been controversial[17-24]. However, several recent studies have suggested that adequate oxygenation by CPAP therapy during sleep for longer than 3 mo is associated with improvements in liver enzymes[25-31], steatosis[20], and fibrosis[28]. Adherence to CPAP therapy could also be a key point in the effectiveness against liver injury in NAFLD with OSA [25,28,30,31].

Liver biopsy is the gold-standard method for diagnosing and assessing severity of hepatic steatosis and for staging fibrosis in NAFLD. However, this method is invasive and associated with pain and a small risk of serious complications. Instead, liver stiffness measurement (LSM) using transient elastography (TE) is a non-invasive, simple-to-perform, totally innocuous, and reproducible method to assess liver fibrosis, and has been demonstrated to be effective in NAFLD. On the other hand, a non-invasive method for assessing hepatic steatosis has also been needed. As fat affects ultrasound propagation, a novel attenuation parameter has been developed to detect and quantify steatosis. This parameter, controlled attenuation parameter (CAP), is based on the ultrasonic properties of the reflected radiofrequency signals acquired by TE using the FibroScan^® M probe (3.5 MHz; Echosens, Paris, France)[32]. Although many reports have demonstrated the efficacy of CPAP therapy for liver enzyme, hepatic steatosis, and fibrosis in NAFLD with OSA, evaluations by LSM and CAP using the FibroScan^® are not well known. We therefore undertook the present study to evaluate the efficacy of adequate effective CPAP therapy for liver injury in NAFLD with OSA by LSM and CAP using FibroScan^® along with serological parameters.

To the best of our knowledge, this is the first prospective cohort study to investigate the effects of reoxygenation by effective CPAP therapy for OSA patients with NAFLD using TE by FibroScan^®. This study demonstrated that effective CPAP therapy for 6 mo in OSA patients with NAFLD could significantly improve serum transaminase activities independent of BW changes. However, no changes were seen in liver stiffness or steatosis from TE or serum fibrosis markers, or in insulin resistance, hs-CRP, or other biochemical parameters.

We first compared anthropometric data, serum biochemical parameters, serum fibrosis markers, and TE measurements between OSA patients of the NAFLD and Normal groups before CPAP therapy. In the analysis of HOMA-IR, NAFLD with OSA represented high insulin resistance as well as NALFD without OSA[35]. TE measurement showed that LSM and CAP measurements in the NAFLD group were significantly higher than those in the Normal group. We also examined hs-CRP in OSA patients of the two groups before CPAP therapy. The concentration of hs-CRP is known to offer a marker of systemic inflammation, and a low-level elevation of hs-CRP level is known to be seen in metabolic syndrome, cardiovascular disease, type 2 diabetes, hypertension[36], some cancers[37], and NAFLD[38]. In this study, hs-CRP was significantly higher in OSA patients with NAFLD than in OSA patients without NAFLD, presumably reflecting the systemic inflammation associated with NAFLD.

In 50 OSA patients with NAFLD, we compared data between before and after 6 mo of CPAP therapy, and both AST and ALT levels were significantly decreased after therapy compared to before. However, LSM and CAP measurement, and other serum biochemical parameters including serum fibrosis markers, were unchanged after therapy. On the other hand, BW tended to decrease after CPAP therapy (P = 0.076). BW loss is known as the most important factor in improving abnormal transaminases for NAFLD[39]. We then considered that improvement of abnormal transaminases in OSA patients with NAFLD may depend more on BW loss than on the reoxygenation achieved by effective CPAP therapy for 6 mo. We therefore next performed uni- and multivariate regression analyses of factors associated with improved ALT level, particularly for the factor of compliance with CPAP therapy.

Compliance with CPAP therapy is one of the most influential factors for the effectiveness of the therapy and is an indispensable factor in this study. Good compliance leads to adequate reoxygenation, but several factors can exert deleterious influences on compliance. The most common reason for poor compliance cited by patients is a troublesome experience using CPAP every night[40]. Although compliance with CPAP therapy is important, standardized evaluations remain lacking. Several studies have used the term “adherence” as a definition of good compliance with CPAP therapy. Kribbs et al[41] and Salepci et al[42] defined adherence as > 4 h per night for > 70% of monitored days, and this definition has been used in many studies. Many prior studies have also expressed adherence to CPAP therapy as mean ± SD for hours per night[18,21,22]. However, detail of the calculations used for this parameter have not been well described. This study therefore originally quantified and expressed compliance with CPAP therapy as a mean usage rate of therapy over 6 mo using a new value, m-CI. In the multivariate regression model, the factor of m-CI for 6 mo of CPAP therapy tended to be associated with improvements in ALT level (P = 0.051), indicating a high probability that compliance with CPAP therapy was associated with improved ALT level independent of BW changes during therapy. The number of participants in this study was so small that the only adjusted factor in multivariate analysis was change in BW.

In addition, to confirm the effectiveness of adequate CPAP therapy for NAFLD with OSA independent of BW changes, we conducted a comparison of biochemical parameters and FibroScan^® data between before and after 6 mo of CPAP therapy in 17 OSA patients with NAFLD who met the following three criteria: No BMI change (BMI < ± 1 kg/m²) in 6 mo; high ALT level (ALT ≥ 30 U/L) before therapy; and good CPAP therapy for 6 mo (m-CI > 0.5). Both AST and ALT levels were significantly decreased after 6 mo of CPAP therapy. However, LSM and CAP measurements including serum fibrosis markers and other biochemical parameters were unchanged. This suggests that reoxygenation by effective CPAP therapy for OSA patients with NAFLD first might improve hepatitis, but does not first improve liver steatosis or fibrosis. Moreover, hs-CRP representing systemic inflammation was also unchanged between before and after CPAP therapy. As levels of hs-CRP were too low, there might be no significant difference in this study. Other systemic inflammatory markers should be considered.

Many studies have investigated the impacts of CPAP therapy on NAFLD with OSA. Kohler et al[18] and Sivam et al[21] investigated the influences on liver enzymes of NAFLD by CPAP therapies for relatively short durations in randomized controlled trials (RCTs). However, no significant improvements in AST or ALT levels were seen compared with controls in either study. On the other hand, several cohort studies have demonstrated improvements in transaminase levels among NAFLD patients with OSA following CPAP therapy for longer than 3 mo[17,19,26-28]. Likewise, a recent meta-analysis indicated that CPAP therapy was associated with significant decreases in both AST and ALT levels among OSA patients, and was more effective in OSA patients receiving treatment for > 3 mo[25]. Similarly, the present study demonstrated that 6 mo of CPAP therapy improved both AST and ALT levels. An RCT to investigate the impacts on NAFLD in OSA patients on CPAP therapy for > 3 mo is warranted.

The influences on liver steatosis and fibrosis of CPAP therapy remain controversial in patients with NAFLD and OSA. Many techniques can be used to evaluate liver steatosis. Sivam et al[21] compared the percentage of intrahepatic lipid by liver volume using magnetic resonance spectroscopy between 2 mo of therapeutic CPAP therapy and 2 mo of sham therapy, finding no significant difference[21]. Moreover, short-term (< 6 mo) CPAP therapy in other studies did not decrease liver fat as evaluated by liver attenuation values on computed tomography (CT)[43,44]. Shpirer et al[19] also evaluated liver steatosis by liver attenuation index from CT, but demonstrated a significant improvement of liver steatosis with long-term (2-3 years) CPAP therapy[19]. In the present study, we evaluated liver steatosis by CAP measurements of TE using FibroScan^®, finding that 6 mo of CPAP therapy did not significantly improve liver steatosis. Improvement of liver steatosis from NAFLD in OSA patients may require over 2 years of effective CPAP therapy. As for liver fibrosis, evaluations can include serum markers, liver fibrosis indexes, and imaging. Buttacavoli et al[20] investigated liver fibrosis by LSM from TE using FibroScan^®, the same method as used in our study, and compared values between before and after 6–12 mo of CPAP therapy in 11 OSA patients[20]. However, no significant differences were identified, supporting our results for 6 mo of CPAP. On the other hand, a few reports have demonstrated significant improvement of liver fibrosis from NAFLD by CPAP therapy. Mesarwi et al[45] evaluated liver fibrosis by serum lysyl oxidase (LOX), an enzyme that cross-links collagen and can serve as a biomarker of hepatic fibrosis[46], and demonstrated significant reductions in serum LOX among 8 OSA patients after 3 mo of CPAP therapy[45]. In addition, Wai et al[47] used AST-to-platelet ratio index (APRI)[47] to evaluate liver fibrosis, and assessed adequate CPAP adherence using a cutoff of > 4 h/night > 70% of treatment nights[28]. Significant improvements in AST, ALT, and APRI were seen with 3 mo CPAP therapy (P < 0.01 each)[28]. Moreover, a dose-response relationship seemed to be present, with patients showing good adherence to CPAP therapy also displaying significantly larger decreases in AST and ALT levels than patients with poor adherence (P < 0.01). Multivariable logistic regression analysis identified CPAP therapy with good adherence (odds ratio 3.93, 95% confidence interval 1.29–11.94) as an independent predictor of NAFLD regression after adjusting for obesity class and severity of OSA[28]. Effective CPAP therapy for an adequate duration was associated with significant biochemical improvement and reductions in CIH-related liver fibrosis, suggesting that reoxygenation over the long term could reduce liver fibrosis from NAFLD only in terms of serum markers. However, to the best our knowledge, no studies have demonstrated patent improvement of liver fibrosis on either visual images or pathological specimens. Further studies are needed to clarify the impacts of reoxygenation on liver fibrosis of NAFLD with OSA.

Several limitations to this study should be considered. The first and biggest limitation of this prospective cohort study was the small sample size, particularly for OSA patients with NAFLD. Only 50 OSA patients in the NAFLD group were investigated by multivariate analysis regarding improvement of ALT levels. Only BW change could be adjusted for due to this small sample size. Second, this study did not include untreated samples and could not compare variations in ALT levels between OSA patients with NAFLD showing good compliance with CPAP therapy and patients with no therapy. Setting patients with moderate to severe OSA to receive no therapy as negative controls would be difficult and ethically questionable. Third, NAFLD in this study was not defined histopathologically using the gold standard of liver biopsy, and the definition was somewhat obscure and determined by the existence of fatty liver images on US and by alcohol consumption. Fourth, our investigation could only be conducted for 6 mo. This was because if OSA patients in our sleep laboratory showed improvements in general condition (e.g., decreased BW and fat around the neck) and AHI after 6 mo of CPAP therapy, therapy would often be changed to combination therapies with an oral appliance. Combination therapy could not be evaluated for adequate reoxygenation and was thus not included in this study.

In conclusion, our findings suggest that transaminase activity in NAFLD patients with OSA could be associated with CIH due to OSA, independent of BW changes, and adequate reoxygenation by effective CPAP therapy for an extended duration of at least 6 mo may thus improve activity of NAFLD. A period of years might be required to improve liver steatosis and fibrosis following adequate reoxygenation. The CIH of OSA might just be part of the cause of NAFLD. However, what kind of NAFLD patients with OSA are indicated for effective CPAP therapy to improve activity of NAFLD has yet to be clarified. Progression from NAFLD to NASH could be preventable if patients with OSA and NAFLD can be appropriately selected and receive effective CPAP therapy. More detailed analyses including experimental research are required. We propose that all NAFLD patients should be analyzed by PSG for the diagnosis of OSA if any symptoms of OSA are apparent.

The authors wish to thank the members of examination laboratory at The Nippon Dental University Medical Hospital for their excellent technical support.