Non-alcoholic fatty liver disease frequency and associated factors at admission of acute stroke

Non-alcoholic fatty liver disease (NAFLD) is emerging as one of the most common chronic liver diseases. Metabolic syndrome is a strong predictor of NAFLD that frequently occurs with obesity, type 2 diabetes mellitus, dyslipidemia, and hypertension [1, 2]. High body mass index (BMI), high triglyceride level, low high-density lipoprotein cholesterol (HDL) level, and high glycated hemoglobin (HbA1c) level strongly suggest the presence of NAFLD [3]. NAFLD has been reported to be a predictor of cardiovascular disease and stroke [2, 4]. Liver enzymes, such as aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyltransferase (GGT) have been associated with the development of cardiovascular disease and all-cause mortality [5]. AST/ALT ratio is a significant predictor of chronic heart failure with reduced left ventricular ejection fraction and cardiovascular mortality [6, 7].

NAFLD may predispose one to ischemic and hemorrhagic stroke [8, 9]. A cross-sectional study conducted from 2011 to 2012 in the general Japanese population demonstrated that the prevalence of NAFLD defined by ultrasonography (US) was 26.4% [10]. Previous retrospective studies reported that the prevalence of NAFLD defined by increased AST or ALT was 7.7% or 42.5%, respectively, in patients with ischemic stroke [11, 12]; however, the prevalence of NAFLD diagnosed by US in patients with acute stroke is unknown. Adults aged 40–79 years have higher risks of metabolic syndrome and cardiovascular disease than individuals of other ages [13], whereas the etiology in stroke patients aged < 40 or ≥ 80 years is frequently different from that in patients aged 40–79 years, e.g., autoimmune disease, atrial fibrillation, or cancer-related thrombosis [14, 15]. Stroke should be aggressively prevented in adults aged 40–79 years with metabolic syndrome or NAFLD. Dietary triglycerides (TG), which are derived from meat, fish, and vegetable oil, influence serum TG and fatty acids (FA) levels. TG, a predictor of NAFLD [10], comprises glycerol and three FAs; thus, the quality of TG depends on FAs. Elevated serum palmitic acid (PA) and oleic acid levels are associated with an increased frequency of lacunar stroke, while elevated serum docosahexaenoic acid (DHA) and arachidonic acid (AA) levels are associated with a decreased incidence of ischemic stroke [16]. Elevated serum dihomo-gamma-linolenic acid (DGLA) level and decreased DHA weight percentage (%) of total fatty acids are associated with the onset of acute ischemic stroke at a younger age [17]. In addition, increased DGLA level and decreased eicosapentaenoic acid (EPA) percentage are associated with the onset of intracerebral hemorrhage (ICH) in young patients [18]. Dyslipidemia is related to NAFLD, and serum FAs are associated with the age of stroke onset [1, 2, 17, 18]; therefore, serum FAs may be associated with NAFLD in patients with stroke. In addition, BMI, triglycerides, type 2 diabetes mellitus, AST/ALT ratio, GGT, and blood pressure may be associated with NAFLD in patients with stroke. If the frequency of NAFLD diagnosed by US is very high in patients with acute ischemic or hemorrhagic stroke, NAFLD may be a critical risk factor for acute stroke. Therefore, identifying the factors associated with NAFLD in patients with acute stroke may lead to stroke prevention. The purpose of this retrospective cross-sectional study was to investigate whether the frequency of NAFLD diagnosed by US was high and identify the factors associated with NAFLD in patients aged 40–79 years with acute stroke.

We included patients with stroke who were admitted to our institution from May 2016 to July 2019. We excluded patients who (1) were younger than 40 years or older than 79 years; (2) had an onset-to-door time of > 24 h; (3) did not undergo body weight measurement; (4) did not undergo abdominal ultrasonography within 5 days of onset of symptom; (5) did not undergo blood examination for serum fatty acids at admission; or (6) had alcoholic fatty liver disease.

Among 1712 patients with stroke admitted to our institution during the study period, 676 patients met our inclusion criteria for analysis (Fig. 1). Patient characteristics are summarized in Supplemental Tables 1 and 2. NAFLD was diagnosed in 267 patients (39.5%). Compared to patients in the NAFLD absence group, patients in the NAFLD presence group were young, had high anthropometric values, had high levels of liver enzymes, had a low AST/ALT ratio, had high levels of serum albumin and HbA1c, had high levels of serum lipids, excluding LDL, and had high BP (Table 1). Further, patients in the NAFLD presence group had high levels of serum fatty acids, except for EPA; high fatty acid% of lauric acid, myristic acid, PA, oleic acid, and DGLA; and low fatty acid% of stearic acid, linoleic acid, and arachidonic acid (Tables 1 and 2). There were no differences in proportions of ischemic stroke or history of statin use between the two groups (Table 1). After evaluating variables with mutually strong correlations among 36 variables with significant differences between the 2 groups (Tables 1 and 2, Supplemental Tables 3 to 12), 17 variables were adopted as variable candidates for multiple logistic regression, and VIF < 3 in the 17 variable candidates was confirmed (Supplemental Table 13). Multiple logistic regression analysis using the 17 variables showed that BMI, AST/ALT ratio, albumin, PA%, and DGLA were independent NAFLD-presence variables (Table 3). Receiver operating characteristic curves demonstrated that the threshold values of BMI, AST/ALT ratio, albumin, PA%, and DGLA for NAFLD presence were ≥ 23.7 kg/m², ≤ 1.25, ≥ 41 g/L, ≥ 24.5%, and ≥ 97.1 μmol/L, respectively (Table 4). These parameters (BMI, AST/ALT ratio, albumin, PA%, and DGLA) might be NAFLD-presence risk factors at their threshold values. Of the 72 patients with no NAFLD-presence risk factors, only 5 patients (6.9%) had NAFLD. Of the 570 patients with 1 to 4 NAFLD-presence risk factors, 232 patients had NAFLD (40.7%; odds ratio, 9.2). Of the 34 patients with all 5 NAFLD-presence risk factors, 30 patients had NAFLD (88.2%; odds ratio, 100.5).

Our results demonstrate that the US-diagnosed NAFLD frequency in stroke patients aged 40 − 79 years was higher than that in the general Japanese population (39.5% vs. 26.4%, respectively), and the factors associated with NAFLD presence were high BMI, low AST/ALT ratio, high albumin level, high PA%, and high DGLA level. The NAFLD frequency was 88% in acute stroke patients with the five risk factors.

In the present study, the NAFLD frequency was very high in the middle-aged and older acute stroke patients compared to the NAFLD prevalence of 26.4% in the general Japanese population [10]. Blood pressure, glucose level, HbA1c level, total cholesterol, and triglycerides were not independent factors of NAFLD presence in acute stroke patients. Among anthropometric variables, only BMI was an independent factor of NAFLD presence in acute stroke patients. Among liver enzymes, only the AST/ALT ratio was an independent factor. The low AST/ALT ratio was also an independent factor for high BMI. Among variables of serum lipids and fatty acids, neither total cholesterol nor triglycerides, but PA%, a saturated fatty acid, and DGLA, a member of the n-6 PUFA family, were independent NAFLD-presence factors in acute stroke patients. Eicosanoid metabolites of DGLA, such as prostaglandin E1 or thromboxane A1, have anti-inflammatory effects [21]. However, DGLA is metabolized to AA, which is metabolized to prostaglandin E2, thromboxane A2, or leukotriene C4 [22]; these have inflammatory effects and increase the risk of acute myocardial infarction [22]. High serum DGLA level was associated with the onset of acute ischemic stroke and ICH at a young age [17, 18] and was also associated with obesity, body fat accumulation, high ALT level, and insulin resistance in patients with type 2 diabetes [23]. High serum DGLA level was associated with not only acute stroke but also NAFLD presence. In a previous study, low fish intake was associated with high DGLA% [24]. The dietary effect of DGLA may be more profound than that of gamma-linolenic acid, which is enzymatically elongated to DGLA [25]. Therefore, diets that contain large amounts of DGLA may increase serum DGLA levels. For example, Japanese beef cattle or crossbred (Holstein × Japanese black) beef steer contains more DGLA than imported beef or lamb [26]. However, it has not been investigated whether serum DGLA levels increase in proportion to the consumption of these proteins and whether high DGLA levels cause obesity and NAFLD.

In Japan, the National Health and Nutrition Survey reported that mean BMI slightly increased in men and women from 1995 to 2016 and the total energy intake and energy intake from proteins decreased, whereas energy intake from fats increased from 1995 to 2016 [27]. Obesity is a stubbornly obvious target for stroke prevention [28]. Therefore, a study was conducted to investigate whether the Japanese diet, characterized by high seafood and plant food composition, affects fatty acid intake and serum fatty acid% [29]. It was found that the Japanese diet, which comprises fish, soybeans, soy products, seaweed, mushroom, konjac, and unrefined cereals, with low amounts of animal fat, meat, poultry with fat, and sweets, including deserts and snacks, was effective in increasing serum n-3 PUFA% and decreasing serum n-6 PUFA% [29]. Serum DGLA% had a strong positive correlation with serum DGLA level (Supplemental Table 8). Therefore, the Japanese diet may decrease serum DGLA level as a member of the n-6 PUFA family. The Japanese diet successfully decreased BW, BMI, and systolic and diastolic blood pressure [30]. If serum DGLA level is decreased, the incidence of NAFLD and obesity is reduced, and the incidence of an acute stroke may be prevented.

In acute stroke patients aged 40 − 79 years, the NAFLD frequency of 39.5% at acute stroke admission was very high. The factors associated with NAFLD-presence were high BMI, low AST/ALT ratio, high albumin level, high PA% level, and high DGLA level. A further study is warranted to determine the effects of the NAFLD-presence factors on stroke onset or prevention.