AIM: To evaluate the effect of resistance training (RT) on non alcoholic liver disease (NAFLD) patients.

METHODS: A randomized clinical trial enrolling NAFLD patients without secondary liver disease (e.g., without hepatitis B virus, hepatitis C virus or excessive alcohol consumption). Patients were randomly allocated either to RT, three times weekly, for 3 mo or a control arm consisting of home stretching. The RT included leg press, chest press, seated rowing, latissimus pull down etc. with 8-12 repetitions, 3 sets for each exercise, for a total duration of 40 min. Hepatic ultrasound, fasting blood tests, anthropometrics and body composition by dual energy X-ray absorptiometry were assessed. At baseline and follow-up, patients filled out a detailed semi-quantitative food frequency questionnaire reporting their habitual nutritional intake. Steatosis was quantified by the hepatorenal-ultrasound index (HRI) representing the ratio between the brightness level of the liver and the right kidney. The HRI has been previously demonstrated to be highly reproducible and was validated against liver biopsy and proton magnetic resonance spectroscopy.

RESULTS: Eighty two patients with primary NAFLD were randomized to receive 3 mo of either RT or stretching. After dropout or exclusion from analysis because of protocol violation (weight change > 3 kg), thirty three patients in the RT arm and 31 in the stretching arm completed the study per protocol. All baseline characteristics were similar for the two treatment groups with respect to demographics, anthropometrics and body composition, blood tests and liver steatosis on imaging. HRI score was reduced significantly in the RT arm as compared to the stretching arm (-0.25 ± 0.37 vs -0.05 ± 0.28, P = 0.017). The RT arm had a significantly higher reduction in total, trunk and android fat with increase in lean body mass. There was no correlation between the reduction in HRI in the RT arm and weight change during the study, but it was positively correlated with the change in trunk fat (r = 0.37, P = 0.048). The RT arm had a significant reduction in serum ferritin and total cholesterol. There was no significant difference between arms in dietary changes and these did not correlate with HRI change.

CONCLUSION: Three months RT improves hepatic fat content accompanied by favorable changes in body composition and ferritin. RT may serve as a complement to treatment of NAFLD.

Lifestyle modifications, including weight reduction and physical activity, improve many of the risk factors for non alcoholic liver disease (NAFLD)[1] and have become the primary treatment modalities for the disease[2]. The role of physical activity (PA) as a potential treatment for NAFLD has been tested in several observational studies and a few clinical trials, mostly testing the effect of aerobic training. Resistance training (strength training) is a means for developing and maintaining muscular strength, endurance, power, and muscle mass that has grown in popularity over the past two decades[3,4]. In a study of the general population, resistance training (RT) was inversely associated with NAFLD. This association remained significant after adjusting to multiple confounders including body mass index (BMI), homeostasis model assessment (HOMA), nutritional factors, adiponectin, and resistin[5]. Two small trials found beneficial effects for RT as a single treatment in NAFLD patients, but results regarding reduction in steatosis were conflicting[6,7]. In a randomized clinical trial (RCT) in 19 sedentary adult NAFLD patients, 8 wk of RT, consisting of 45 min sessions trice weekly, led to a reduction in liver fat without weight loss[7]. In an uncontrolled clinical trial in 12 obese adolescents, a three-month RT program consisting of 1 h sessions twice weekly did not change hepatic fat content but improved hepatic insulin sensitivity[6]. In a recent RCT among type-2 diabetic patients, it was demonstrated that resistance training and aerobic training are equally effective in reducing hepatic fat content[8].

NAFLD patients report a poorer health-related quality of life compared with healthy United States population both on physical and mental health scores[9]. Furthermore, fatigue is a common symptom in NAFLD patients[10], and they report low scores for vitality[9]. Although NAFLD patients understand the benefits of exercise, they lack the confidence to perform it and express a fear of falling[11]. The potential benefits of RT are not only to cardiovascular health and to weight management but it also improves balance and reduces the risk of falls[12-14]. RT improves several components of physical function, contributes to health-related quality of life[15,16], and is well tolerated even among patients with coronary heart disease or the elderly[15,17]. In recent years, increasing attention has been paid to RT as a useful adjunctive tool of exercise in various metabolic diseases including diabetes and heart disease[4,18,19]. Indeed, the American Heart association and American College of Sport Medicine (ACSM) recommend RT at least twice a week in addition to aerobic training[4]. For those patients who may have physical limitation or low motivation that prevents them from performing aerobic PA, RT can serve as an alternative option.

Therefore we conducted a randomized controlled trial of RT vs stretching. Our main aim was to evaluate the effect of 3 mo RT on the presence of fatty liver measured by abdominal ultrasound and on the hepato-renal index (HRI) as a quantitative objective measurement of steatosis. We also evaluated the effect of the RT program on liver enzymes, metabolic parameters and body composition.

We conducted a RCT (sealed envelopes randomization stratified by gender) in consecutive patients with ultrasound diagnosed fatty liver attending the liver clinic at the Tel-Aviv Medical center during 2010-2012 and community regional HMO’s clinics. Inclusion criteria were age between 20-65 years and a diagnosis of fatty liver by ultrosound in the past 6 mo and on the baseline US examination. Exclusion criteria were any known secondary liver disease including the presence of hepatitis B surface antigen or anti-hepatitis C virus antibodies, excessive alcohol consumption defined as ≥ 30 g/d in men or 20 g/d in women, administration of medical treatment that may elevate alanine aminotransferase (ALT) or lead to hepatic steatosis, known diabetes, major chronic diseases including: renal, cardiovascular, lung, uncontrolled hypertension, inflammatory bowel disease, active cancer, autoimmune disorders and orthopedic contraindications for RT. Adults with diabetes were excluded to avoid a confounding effect, since it is unclear whether they would have the same response to physical training and since changes in antidiabetic medications during the trial might occur. We also excluded patients regularly performing RT in the 3 mo or 6 mo prior to study enrolment for novice and progressive trainee respectively (novice-trained continuously less than 2 mo, progressive-trained continuously more than 2 mo). Patients performing vigorous aerobic PA in the 3 mo prior to the study, defined as aerobic exercise ≥ 5 d a week 30 min at moderate pace or 3 d a week 20 min at vigorous pace or 4 times a week with combination of both and patients with recent weight reduction (more than 3 kg in the last 3 mo) were also excluded.

A sample size of 32 patients in each group was calculated to be needed for a 90% power to detect a difference of 0.25 with a standard deviation of 0.30, based on previously published data on HRI change following weight change[20], with a 0.050 two-sided significance level. Additional 20% patients were recruited taking into consideration attrition or protocol violation.

The study was approved by the Tel-Aviv Medical center ethics committee and all patients signed an informed consent. The study was pre-registered in the NIH registration website (TRIAL no. NCT01264198).

In this randomized controlled trial, NAFLD patients without diabetes underwent either RT trice weekly or static stretching. The results suggest that RT exerts beneficial effects on several clinical and biochemical parameters including liver fat and body composition.

The 2007 update of the American Heart Association dealing with resistance exercise concludes that RT should be viewed as a complement to aerobic exercise[14]. However, the beneficial effect of RT for patients with steatosis was so far not supported by strong evidence.

The present trial is one of the first to test this question on a large group of patients. We have demonstrated a significant reduction in steatosis as measured by an objective ultrasonographic tool, the HRI. The modest relative reduction of about 10% is similar to the 13% relative reduction in steatosis previously demonstrated[7]. Although the RT arm had a nonsignificant reduction in caloric intake and a small but significant weight reduction of less than half a kg, these changes did not correlate with HRI change. Our results suggest that the reduction in steatosis in the RT arm cannot be explained by weight loss or dietary change. RT also improved body composition, most importantly trunk fat mass that was positively correlated with the change in HRI.

The weight-reduction independent beneficial effect of aerobic exercise in NAFLD is supported by clinical trials demonstrating a relative reduction of hepatic triglyceride concentration by 21%-35%[33-35] following supervised training such as cycling. However, in a trial of a more modest activity that included brisk walk, there was a relative reduction of 10.3% in liver fat[36], similar to the one observed by the present study. In previous published trials about the effect of RT in adult NAFLD patients, there was a significant improvement in glycemic control and no improvement in liver enzymes[7,8]. Our study did not demonstrate improved glucose metabolism, this discrepancy may stem from the exclusion of diabetic patients from our study. It was previously shown that RT improves hyperglycemia, only in patients with disturbed glucose metabolism or diabetes[37-41]. Our study showed significant improvement in the liver enzymes (ALT and aspartate aminotransferase-AST) within group but with no difference between arms. Interestingly, there seems to be a limited correlation between exercise and liver enzyme reduction. Aerobic PA led to a significant reduction in liver enzymes in some trials[42] while no reduction was seen in other trials despite reduced steatosis[8,34,35]. With specific reference for resistance training, it has been shown that weightlifting exercise resulted in increases in liver enzymes; AST and ALT, though the underlying mechanisms are unknown[43]. Thus, it may be in our study that the reduction of liver enzymes in the RT arm was masked and is underestimated.

Serum ferritin was significantly reduced only in the RT group. This novel effect of RT in NAFLD patients, that has been previously demonstrated with aerobic training[42] and lifestyle intervention[44], is of importance due to the strong association of ferritin with fibrosis and inflammation[45,46] and with insulin resistance (IR)[23,47] in NAFLD patients. The mechanisms by which PA, and specifically RT, reduces serum ferrtin are unknown. Serum ferritin is associated with insulin resistance[23,48] and is an acute phase protein that can be induced in the setting of systemic or hepatic inflammation[49-52]. Serum ferritin was demonstrated to be a predictor of histologic severity including steatosis[46]. Thus, we can assume that a reduction in liver fat demonstrated in our study or improved hepatic insulin sensitivity by RT, that was demonstrated in another study[6], led to a reduction of serum ferritin. Furthermore, RT may have an anti-inflammatory effect, as demonstrated by increasing adiponectin levels[53], and serum ferritin was found to be inversely correlated with serum adiponectin[54]. However, it is unclear if hyperferritinemia in NAFLD is simply a consequence of disease severity or actively contributes to disease progression[46]. Ferritin was found to inhibit the secretion of apolipoprotein B and in this way may alter cholesterol and triglyceride transport in the liver[55]. Interestingly, some studies reported that PA plays an important role in reducing serum ferritin concentration[56,57] and this may be another explanation for the reduction in liver fat in our study.

Another beneficial effect of RT in our study was a significant reduction of serum cholesterol. Although data regarding the effect of RT on lipid metabolism are equivocal, reduction of serum total cholesterol and LDL by resistance training has been previously demonstrated in a meta-analysis of randomized controlled trials[58]. It is well established that liver steatosis is associated with IR and lipid abnormalities including alteration in cholesterol metabolism[59-62]. Recent data show that increased IR contributes to the shift in cholesterol metabolism to increased synthesis and decreased absorption, independent of body weight[63-65]. Several studies have demonstrated that resistance training improves IR, including hepatic IR[6,66], and therefore may contribute to decreased synthesis of hepatic cholesterol. However, the precise mechanisms involved still need to be clarified[67].

Professional societies recommend ≥ 30 min of moderate-intensity aerobic PA on most, and preferably all, days of the week, or vigorous-intensity PA ≥ 3 times per week for ≥ 20 min each time. However, only 27.7% United States adults meet recommended levels of either moderate or vigorous physical activity, whereas 29.2% report no regular PA outside of their work[68,69]. Moreover, the prevalence of physically active adults among patients with diabetes is lower than in those without diabetes[70] and they are less likely to meet PA recommendations[71]. In NAFLD patients, compliance may be even lower because fatigue has been demonstrated to be markedly higher in NAFLD patients compared to controls, and is associated with inactivity and excessive daytime sleepiness[10]. Therefore, an alternative or a complement form of exercise that may be easier to perform or to adhere to, such as RT, may be helpful in the treatment of NAFLD patients.

The major limitation of this study is that due to ethical and practical considerations we were unable to perform repeated liver biopsies in our short term study preventing any inference regarding the effect of RT on inflammation and fibrosis. Ideally, quantification of liver fat is performed by liver biopsy or magnetic resonance spectroscopy (¹H-MRS) that is an accepted noninvasive method to reliably quantify steatosis[72,73].

HRI has been validated vs liver biopsy and provides a highly sensitive, objective and quantitative tool for liver fat evaluation with a high correlation (r = 0.82, P < 0.001) and a kappa of 0.75 as compared with histological steatosis[26]. Another group compared HRI to ¹H-MRS as a reference standard, demonstrating high correlation (r² = 0.92, P < 0.0001) and therefore confirming that HRI can be a valuable analytic tool in clinical investigation[72]. It was also recently shown that HRI highly correlates with biochemical surrogate markers of liver steatosis: the fatty liver index (FLI) (r = 0.55, P < 0.001) and the SteatoTest (r = 0.52, P < 0.001)[74]. Furthermore, since a baseline biopsy wasn’t performed, we were unable to distinguish between patients with simple steatosis or NASH. NASH is the NAFLD variant that needs to be treated more urgently due to a worse natural history of progression to cirrhosis, hepatocellular carcinoma[75] and increased liver related mortality[76]. Recently, ultrasonographic FLI, another semi-quantitative evaluation tool of hepatic steatosis, was demonstrated to successfully predict biopsy diagnosed NASH[77].

In conclusion, this relatively large randomized clinical trial demonstrated a significant reduction in steatosis, as assessed by HRI, during 3 mo RT accompanied by favorable changes in body composition and reduction of serum ferritin.

We thank the "Great Shape" gyms for their cooperation and contribution during the study.