Department of Hepatology and Pancreatology, Kawasaki Medical School, Kurashiki, Okayama, Japan
Correspondence: Masaaki Korenaga, The Research Center for Hepatitis and Immunology, National Center for Global Health and Medicine at Kohonodai, 1-7-1 Kohnodai, Ichikawa, Chiba 272-8516, Japan. Email; firstname.lastname@example.org
Little is known about the effects of non-alcoholic fatty liver disease (NAFLD) on energy metabolism, although this disease is associated with metabolic syndrome. We measured non-protein respiratory quotient (npRQ) using indirect calorimetry, which reflects glucose oxidation, and compared this value with histological disease severity in NAFLD patients.
Subjects were 32 patients who were diagnosed with NAFLD histopathologically. Subjects underwent body composition analysis and indirect calorimetry, and npRQ was calculated. An oral glucose tolerance test was performed, and plasma glucose area under the curve (AUC glucose) was calculated.
There were no differences in body mass index, body fat percentage or visceral fat area among fibrosis stage groups. As fibrosis progressed, npRQ significantly decreased (stage 0, 0.895 ± 0.068; stage 1, 0.869 ± 0.067; stage 2, 0.808 ± 0.046; stage 3, 0.798 ± 0.026; P < 0.005). Glucose intolerance worsened and insulin resistance increased with fibrosis stage. npRQ was negatively correlated with AUC glucose (R = −0.6308, P < 0.001), Homeostasis Model of Assessment – Insulin Resistance (R = −0.5045, P < 0.005), fasting glucose (R = −0.4585, P < 0.01) and insulin levels (R = −0.4431, P < 0.05), suggesting that decreased npRQ may reflect impaired glucose tolerance due to insulin resistance, which was associated with fibrosis progression. Estimation of fibrosis stage using npRQ was as accurate as several previously established scoring systems using receiver–operator curve analysis.
npRQ was significantly decreased in patients with advanced NAFLD. Our data suggest that measurement of npRQ is useful for the estimation of disease severity in NAFLD patients.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Non-alcoholic fatty liver disease (NAFLD) is one of the most common chronic liver diseases. NAFLD is associated with metabolic syndrome and insulin resistance and often involves abnormal glucose and lipid metabolism.[1-6] Based on this, the NAFLD Asia–Pacific Working Party has recommended screening for metabolic syndrome and body composition in all NAFLD patients. NAFLD treatment consists of diet and exercise interventions for weight loss,[4-6] and nutritional guidance and management are essential. As a part of a nutritional guidance and management program, our institution performs anthropometric measurement of NAFLD patients using a body composition analyzer, evaluation of glucose metabolism using a 75-g oral glucose tolerance test (OGTT), and evaluation of energy metabolism using indirect calorimetry. These basic tests are performed in routine practice.
Indirect calorimetry is a method used in physiological testing and enables easy and non-invasive evaluation of energy metabolism in real time. The non-protein respiratory quotient (npRQ) calculated from indirect calorimetry data represents the ratio of carbohydrate to fat oxidation, and its value is said to be an indicator of prognosis in liver cirrhosis. In addition, although it has been reported that NAFLD disease progression is associated with glucose intolerance[9-12] and visceral fat accumulation, no previous study has examined the specific relationship between NAFLD pathology and these nutritional parameters.
One aim of the present study was to elucidate whether nutritional status, as estimated by indirect calorimetry, 75-g OGTT and body composition analysis, was related to NAFLD disease progression. The other aim was to elucidate whether these parameters were useful for prediction of the severity of disease.
Subjects were 32 patients diagnosed with NAFLD/non-alcoholic steatohepatitis (NASH) by biopsy between April 2009 and March 2011 at our institution. All patients had untreated impaired glucose tolerance (no drug treatment) and present and past alcohol consumption of 20 g or less per week. No patient had been treated with drugs, such as tamoxifen, that can induce NAFLD/NASH. Patients were excluded if they had liver cirrhosis with decreased npRQ accompanied by protein energy malnutrition (PEM). Other exclusion criteria included a history of liver diseases such as primary biliary cirrhosis, autoimmune hepatitis, hepatitis B infection or hepatitis C infection. Hepatocellular carcinoma (HCC) was not detected in any patient.
The study protocol was approved by the institutional review board. Written informed consent was obtained from all patients before trial registration. For all patients, tests were performed in the hospital under resting, fasted conditions in the early morning.
All subjects were hospitalized for at least 2 days to undergo a liver biopsy. Indirect calorimetry and 75-g OGTT were performed before liver biopsy, as described below. Anthropometric measurements and laboratory analysis were carried out before the indirect calorimetry study. All subjects received nutritional guidance from dieticians and were prescribed medical nutrition therapy (energy 25–30 kcal/kg ideal bodyweight).
Physical examination and serum biochemistry
Anthropometric measurements (body mass index [BMI], body fat percentage, and visceral fat area [VFA]) were performed using a body composition analyzer (InBody 720; BIOSPACE, Tokyo, Japan). We previously determined VFA values using a body composition analyzer and performed abdominal computed tomography using Fat Scan software (E2 system, Osaka, Japan) in 27 NAFLD patients. There was a strong correlation between these two modalities (n = 27, R = 0.9319, P < 0.0001; unpubl. data). Venous blood samples were collected in the early morning after patients had fasted for 12 h. These samples were used for several biochemical tests.
NAFIC score, NAFLD fibrosis score and FIB-4 index were calculated using previously reported formulas.
A 75-G OGTT was performed, and plasma glucose and immunoreactive insulin (IRI) were measured at 0, 30, 60, 90 and 120 min after glucose loading. Based on the classification of the Expert Committee on the Diagnosis and Classification of DM, individuals were diagnosed with impaired fasting glucose (IFG) if they had fasting plasma glucose levels of 110 mg/dL or more, but less than 126 mg/dL, and if they had a plasma glucose level less than 140 mg/dL at 120 min after glucose loading. Individuals were diagnosed with impaired glucose tolerance (IGT) if they had plasma glucose levels of less than 110 mg/dL at 0 min after loading and exceeding 140 mg/dL at 120 min after loading. Individuals were diagnosed with diabetes mellitus (DM) if they had plasma glucose levels exceeding 200 mg/dL at 120 min after loading. Homeostasis Model of Assessment – Insulin Resistance (HOMA-IR) was calculated using the following formula: HOMA-IR = fasting insulin (mU/mL) × plasma glucose (mg/dL) / 405. Plasma glucose area under the curve (AUC glucose) and IRI area under the curve (AUC IRI) were calculated using methods reported previously.
Energy metabolism was measured by indirect calorimetry (Aero Monitor AE-300s; Minato Medical Science, Osaka, Japan). A previously reported method was used to measure oxygen uptake and carbon dioxide exhalation under resting, fasted conditions in the early morning. Twenty-four-hour urine nitrogen levels were also measured. The resulting values were used to calculate npRQ and resting energy expenditure (REE). The basal metabolic rate (BMR) was calculated using the Harris–Benedict formula.
All samples were diagnosed by a pathologist who was not notified of subjects' clinical data or course. The classification of Brunt et al. was used for fibrosis staging, and disease activity was assessed using the NAFLD activity score (NAS).
Statistical analysis was performed using SPSS ver. 20.0 software (SPSS, Chicago, IL, USA). Results were expressed as mean ±standard deviation or standard error of the mean. A χ2-test was used for categorical variables. A Student's t-test or Mann–Whitney U-test was used to compare two groups. One-way anova or Kruskal–Wallis analysis followed by a post-hoc test was used to compare multiple independent groups. Correlation was assessed using Spearman's correlation coefficient. Receiver–operator curves (ROC) were used to assess discrimination ability. Statistical significance was defined as P < 0.05.
The clinical and biochemical characteristics of patients enrolled in the study are summarized in Table 1. The 32 subjects (24 male, eight female) had a mean age of 45.4 years (range, 27–75). BMI ranged 22.0–38.8 kg/m2 and averaged 27.2 kg/m2. Serum alanine aminotransferase (ALT) levels ranged 22–200 IU/L and averaged 95.6 IU/L. Histological findings are shown also in Table 1. Fibrosis stages were determined according to Brunt et al.'s classification, and there were eight patients at stage 0, 10 patients at stage 1, seven patients at stage 2 and seven patients at stage 3. For NAS, there were six patients with scores of less than 3, 20 patients with scores of 3 or 4, and six patients with scores of 5 or more.
Table 1. Characteristics of the patient population (n = 32)
Results are expressed as mean ± standard deviation.
aStage and grade on histological assessment were determined using Brunt's classification.
Body mass index, body fat percentage and VFA tended to increase as fibrosis progressed. However, there was no significant difference in these parameters among groups with different Brunt stages (Table 2).
Table 2. Clinical features and laboratory data of NAFLD/NASH patients determined using Brunt et al.'s classification
Four patients (12.5%) had HbA1c levels of at least 6.1% or fasting glucose of at least 110 mg/dL and in whom impaired glucose tolerance was suspected before the 75-g OGTT. The 75-g OGTT did not reveal a normal glucose tolerance pattern in any patient, and all patients had impaired glucose metabolism. One patient (3.1%) had IFG, 16 patients (50.0%) had IGT and 15 patients (46.9%) had DM.
Fasting glucose, HbA1c and AUC glucose increased as fibrosis progressed, and glucose metabolism was significantly worsened with fibrosis progression (Table 2). The 75-g OGTT revealed a correlation between postprandial hyperglycemia and Brunt stage (Fig. 1a). There were significant differences among fibrosis stages in plasma glucose levels at 0, 30 and 120 min after loading (P < 0.05 using Kruskal–Wallis analysis).
In patients with fibrosis stages 1–3, fasting insulin levels were increased and HOMA-IR was elevated, indicating the presence of insulin resistance. This tendency was significantly more pronounced in more advanced fibrosis stages (Table 2). In the 75-g OGTT (Fig. 1b), insulin secretion was delayed and postprandial hyperinsulinemia was observed for all stages compared with healthy controls, as previously reported. In particular, stage 3 patients had significantly greater hyperinsulinemia than patients at the other stages at 0, 90 and 120 min after loading (P < 0.05 using Kruskal–Wallis analysis).
Non-protein respiratory quotient values determined using indirect calorimetry data significantly decreased as fibrosis progressed (Table 2, Fig. 2a). In addition, npRQ values significantly decreased as NAS increased (Fig. 2b). There was no relationship between npRQ and disease activity (Fig. 2c).
Resting energy expenditure and BMR predicted using the Harris–Benedict formula did not differ among Brunt stages or NAS classifications (data not shown). The ratio of REE to BMR (REE/BMR) also did not differ among Brunt stages (Table 2) or NAS classifications (data not shown). In addition, this ratio was within the normal range (0.9 < REE/BMR < 1.1), indicating that most subjects were in normal metabolic states, and not hyper- or hypometabolic states.
There were significant differences among stages in AST, ALT, type IV collagen 7S and ferritin levels (Table 2). However, there were no significant differences among stages with respect to other parameters, including γ-glutamyl transferase (γ-GT), total cholesterol, triglyceride and non-esterified fatty acid (NEFA) (Table 2), AST/ALT ratio, hyaluronic acid, platelet count and prothrombin time (data not shown).
Comparison of patients with mild fibrosis (stages 0–1) and patients with more advanced fibrosis (stages 2–3)
To identify factors correlated with fibrosis in NAFLD patients, we divided subjects into two groups – those with mild fibrosis (stages 0–1) and those with more advanced fibrosis (stages 2–3) – and compared clinical features.
Patients with more advanced fibrosis had significantly higher values than patients with mild fibrosis for the following parameters: VFA, serum AST, ALT, P-III-P, type IV collagen 7S, fasting glucose, fasting insulin, HOMA-IR, HbA1c and AUC glucose (Table 3). Patients having more advanced fibrosis had significantly lower npRQ values than patients with mild fibrosis (Table 3). There were no significant differences between the two groups with respect to other parameters, including γ-GT, total cholesterol, triglyceride and NEFA (data not shown).
Table 3. Comparison of clinical features and laboratory data between stages 0–1 versus stages 2–3 in NAFLD/NASH patients
Differences between two groups were determined using Student's t-test, Mann-Whitney's U-test, or chi-square test.
ALT, alanine aminotransferase; AST, aspartate aminotransferase; AUC glucose, blood glucose area under the curve; AUC IRI, immunoreactive insulin area under the curve; BMI, body mass index; DM, diabetes mellitus; γ-GT, γ-glutamyl transferase; HOMA-IR, Homeostasis Model of Assessment – Insulin Resistance; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; npRQ, non-protein respiratory quotient; 75-g OGTT, 75-g oral glucose tolerance test; VFA, visceral fat area.
42.9 ± 12.35
48.6 ± 11.6
26.3 ± 4.6
28.4 ± 2.8
Percent body fat (%)
29.4 ± 7.6
33.4 ± 6.7
119.7 ± 36.8
153.3 ± 39.0
40.2 ± 19.4
98.2 ± 101.7
70.1 ± 48.2
128.4 ± 65.5
0.53 ± 0.10
0.81 ± 0.38
Type IV collagen 7S (ng/mL)
3.48 ± 0.84
5.64 ± 2.83
Fasting glucose (mg/dL)
87.2 ± 8.1
104.3 ± 18.2
Fasting insulin (μU/mL)
9.9 ± 4.5
21.3 ± 14.2
2.1 ± 1.1
5.2 ± 3.1
Hemoglobin A1c (%)
5.5 ± 0.5
6.0 ± 0.7
AUC glucose (mg·h/dL)
379.4 ± 64.2
473.4 ± 49.4
AUC IRI (μU·h/mL)
254.2 ± 100.6
371.8 ± 261.6
0.881 ± 0.067
0.803 ± 0.036
Correlation of npRQ with parameters of glucose and fat metabolism and body composition
We next compared npRQ to parameters of glucose and fat metabolism and body composition. There was a negative correlation between npRQ and AUC glucose (R = −0.6308, P < 0.001 using Spearman's correlation coefficient) (Fig. 3a). There was also a negative correlation between npRQ and HOMA-IR (R = −0.5045, P < 0.005) (Fig. 3b). A weak negative correlation was found between npRQ and fasting glucose (R = −0.4585, P < 0.01) (Fig. 3c), fasting insulin (R = −0.4431, P < 0.05) (Fig. 3d), γ-GT (R = −0.4428, P < 0.05), plasma glucose levels 120 min after loading (R = −0.3684, P < 0.05) and IRI levels 120 min after loading in the OGTT (R = −0.3772, P < 0.05). No significant correlation was found between npRQ and AUC IRI (R = −0.2992, P = 0.1021), total cholesterol (R = −0.2499, P = 0.1678), triglyceride (R = −0.0617, P = 0.7599), NEFA (R = −0.0629 P = 0.7367), BMI (R = −0.3165, P = 0.0776), body fat percentage (R = −0.1233, P = 0.5088) or VFA (R = −0.2308, P = 0.2199). Based on these results, we speculated that low npRQ in NAFLD is associated with impaired glucose tolerance due to insulin resistance.
Comparison of npRQ to several parameters and previously established scoring systems
We calculated area under the ROC (AUROC) for npRQ and for several of the parameters shown in Table 2. We compared these AUROC to see if they could differentiate stage 3 from stages 0–2, stages 2–3 from stages 0–1, and stage 0 from stages 1–3. Table 4 summarizes these results. For differentiation of stages 3 from stages 0–2, the calculated AUROC was greatest for NAFIC score (0.9200), followed by HOMA-IR (0.9100), type IV collagen 7S (0.8820), AUC glucose and fasting insulin (0.8743), ferritin (0.8690) and npRQ (0.8343). For differentiation of stages 2–3 from stages 0–1, the AUROC for npRQ was greatest (0.8849), followed by HOMA-IR (0.8846), AUC glucose (0.8690), fasting glucose (0.8651), NAFIC score (0.8373) and fasting insulin (0.8234). For differentiation of stage 0 from stages 1–3, the AUROC for ALT was greatest (0.8568), followed by AST (0.8542), fasting insulin (0.8490), HOMA-IR and AUC glucose (both 0.8478), NAFIC score (0.8281) and npRQ (0.8203).
Table 4. AUROC for npRQ, other biochemical parameters and scoring systems for NAFLD/NASH patients
AUROC Stage 0 vs. Stages 1–3
AUROC Stages 0–1 vs. Stages 2–3
AUROC Stages 0–2 vs. Stage 3
ALT, alanine aminotransferase; AST, aspartate aminotransferase; AUC glucose, plasma glucose area under the curve; AUC IRI, immunoreactive insulin area under the curve; AUROC, area under the receiver–operator curve; HOMA-IR, homeostasis model assessment of insulin resistance; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; npRQ, non-protein respiratory quotient.
In each of the three comparisons of stages, AUROC for npRQ, HOMA-IR, AUC glucose, fasting insulin and NAFIC score were all over 0.8000 and showed relatively good results. To differentiate stage 3 from stages 0–2, the AUROC for type IV collagen 7S and ferritin were high; however, AUROC for these parameters were not able to differentiate stage 0 from stages 1–3. This is due to the fact that these two parameters had elevated values in stage 3 and there was no significant difference from stage 0 to stage 2. The AUROC for NAFLD fibrosis score and FIB-4 index were lowest for differentiation of stage 0 from stages 1–3, and increased for differentiation of stages 2–3 from stages 0–1 and differentiation of stage 3 from stages 0–2. This result suggests that these two methods of scoring fibrosis had a relatively high degree of accuracy in distinguishing severe from mild or no fibrosis. AUROC for AST and ALT could be used to differentiate stage 0 from stages 1–3, but were not as accurate in differentiating stage 3 from stages 0–2.
Non-alcoholic fatty liver disease comprises a wide spectrum of conditions ranging from simple steatosis to NASH, which can progress to cirrhosis and HCC. Patients with advanced liver fibrosis are considered to be at high risk for liver failure and HCC.[1-6] Thus, it is important to efficiently identify patients at risk for advanced fibrosis among a large number of NAFLD patients. In addition, differentiation of early-stage NASH allows for early intervention, which can improve patients' outcomes. Liver biopsy is the most reliable method for the diagnosis and determination of fibrosis stage in patients with NASH. However, it is widely acknowledged that biopsy is costly and runs the risk of sampling error and procedure-related morbidity and mortality. Some guidelines recommend that liver biopsy should be considered in patients who are at risk for NASH with advanced fibrosis,[4-6] but this recommendation is not universally accepted. Therefore, various parameters have been proposed as tools to distinguish NASH from NAFLD, or to determine fibrosis stage. Various serum biochemical markers, including indicators of oxidative stress, insulin resistance, inflammation, and apoptosis, have been used for this purpose.[1-6]
With regard to glucose metabolism and insulin resistance, a 75-g OGTT may help clinicians to identify high-risk patients for more intensive monitoring and treatment because blood glucose and insulin levels in the OGTT are important factors for the diagnosis of NAFLD and prediction of fibrosis.[11, 12, 24-26] Studies in which a 75-g OGTT was performed have shown that impaired glucose tolerance is common even in NAFLD patients without overt DM,[11, 12, 24-26] and that postprandial hyperglycemia is associated with advanced fibrosis.[12, 24, 26] Postprandial hyperinsulinemia is also observed in nearly all NAFLD patients, even those with normal glucose tolerance.[11, 12] Kimura et al. reported that postprandial hyperinsulinemia, as indicated by an OGTT, became more marked as fibrosis stage advanced.
Indirect calorimetry provides important information about energy expenditure, npRQ, and the rate of oxidation of three major macronutrients (carbohydrates, fat and protein) based on respiratory gas exchange and urinary nitrogen excretion. Indirect calorimetry is considered the gold standard for assessing energy expenditure and aids in the delivery of the highest quality of nutritional care. The advantages of this modality are that it is non-invasive, portable enough to be done at bedside, easy to operate and inexpensive. Many studies have used this modality to estimate the nutritional state of cirrhotic patients with chronic liver disease. Tajika et al. have reported that low npRQ derived from PEM is associated with survival in patients with viral liver cirrhosis. We have previously used indirect calorimetry in cirrhotic patients to evaluate the effects of nutritional treatment with branched-chain amino acids.[19, 27-29] In diabetic patients, indirect calorimetry is often used to estimate glucose oxidation rate and it has been reported that both glucose oxidation and non-oxidative disposal are impaired during hyperinsulinemic clamping in type 2 DM patients. Although indirect calorimetry is used for assessment in several metabolic diseases, this is the first report to examine indirect calorimetry data from patients with NAFLD. The objective of the present study was to determine how energy metabolism, as estimated by indirect calorimetry, is related to the clinicopathogenesis of NAFLD with glucose intolerance.
We found that npRQ decreased in severity with increased fibrosis stage in NAFLD patients. This observation raised the question of the mechanism underlying decreased npRQ in patients with advanced fibrosis. As fibrosis progressed, npRQ decreased significantly, glucose intolerance worsened and insulin resistance increased (Tables 2, 3). In fact, negative correlations were seen between npRQ and several parameters of glucose intolerance: AUC glucose, HOMA-IR, and fasting glucose and insulin levels. Thus, we speculated that decreased npRQ in NAFLD results from glucose intolerance due to insulin resistance, which worsened with fibrosis stage. Decreased npRQ can reflect reduced glucose oxidation and enhanced lipid oxidation.[8, 19, 27-29] It was reported that peripheral insulin resistance reduces glucose oxidation and glucose uptake in peripheral skeletal muscle. This reduction in glucose uptake may reflect hyperglycemia and decreased glucose oxidation, because the amount of free cellular glucose available for oxidation is reduced. However, it was also speculated that the low glucose oxidation rate seen in viral cirrhosis is a result of reduced glucose production due to decreased hepatic glycogen. In our study, glycogen levels in the liver were not measured directly and thus a definitive statement cannot be made. However, npRQ was low even in the one patient with mild (stage 1) fibrosis, who was found not to be in a state of malnutrition as determined by anthropometry and in whom glycogen storage was likely not decreased to a large extent. Therefore, it is unlikely that the low npRQ in these patients primarily reflects decreased glycogen stores. However, we do not suggest that low npRQ in NAFLD patients is solely due to glucose intolerance because whole-body energy metabolism is a complex process, and thus it is possible that other factors also contribute to low npRQ. Yokoyama et al. have reported that the glucose oxidation rate of subjects with type 2 diabetes is inversely correlated with BMI, body fat percentage and plasma fatty acid levels, suggesting that decreased glucose oxidation and increased fat oxidation may be potently affected by adiposity. In our study, npRQ showed no correlation with lipid parameters, including serum total cholesterol, triglyceride and NEFA, or anthropometric parameters such as body fat percentage and VFA. Thus, low npRQ was speculated to be associated with decreased glucose oxidation due to glucose intolerance, but not with increased fat oxidation, which occurs in the maintenance and development of hyperglycemia during decompensation. Decreased npRQ was correlated with glucose intolerance and fibrosis stage, suggesting that the clinicopathogenesis of NAFLD is closely associated with glucose intolerance, and that early intervention for glucose intolerance is important in clinical practice.
To examine the utility of npRQ as a marker of disease progression in NAFLD, we compared AUROC for npRQ with those for various other parameters and scoring systems in three patterns to discriminate NASH from NAFLD (stage 0 vs stages 1–3), significant fibrosis (stages 0–1 vs stages 2–3) and advanced fibrosis (stages 0–2 vs stage 3). As the decrease in npRQ became significant at stage 2 (Fig. 2a), the AUROC for npRQ for differentiation of stages 2–3 from stages 0–1 was superior to other parameters. Therefore, our results indicate that decreased npRQ can be used to detect NASH, including relatively early stage NASH in many NAFLD patients. In addition to npRQ, AUROC for HOMA-IR, AUC glucose, fasting insulin and NAFIC score were each approximately 0.850 and also showed differences for each of the three differentiation patterns. Therefore, these parameters also have the ability to detect NASH from the early stages to the development of severe fibrosis. NAFIC score, the scoring system for fibrosis proposed by Sumida et al., comprises three measurements (serum ferritin, insulin and type IV collagen 7S) and is easy to calculate. A validation study by the Japan Study Group of NAFLD (JSG-NAFLD) reported that NAFIC score was superior to other several previously established scoring systems in detecting NASH with fibrosis among Japanese NAFLD patients, and also for predicting severe fibrosis. In the present study, AUROC for NAFIC score for differentiation of stage 3 from stages 0–2 was 0.9200 and was the highest among the various parameters, supporting the conclusions of the JSG-NAFLD report. NAFLD fibrosis score, which consists of six variables (age, BMI, hyperglycemia, platelet count, albumin and AST/ALT ratio) has been reported to reliably predict advanced fibrosis. In a meta-analysis by Angulo et al., NAFLD fibrosis score had an AUROC of 0.85 for predicting advanced fibrosis (stages 3–4). The fact that subjects with stage 4 were excluded from the present study but were included as “advanced stage” in previously reported studies,[14, 15, 34] may also contribute to the lower value of AUROC in the present study. In addition, it is uncertain whether a scoring system established using data from Caucasian populations is applicable to Asian patients, because Asian patients tend to develop NASH and other metabolic complications at a lower BMI than Caucasians. The FIB-4 index was developed as a scoring system for estimation of liver fibrosis in subjects with HIV and hepatitis C virus co-infection. It relies on patient age, AST, ALT and platelet count. The advantage of FIB-4 index is that it is easy to calculate and does not require the use of BMI. In a validation study of JSG-NAFLD, FIB-4 index was superior to other fibrosis scoring systems in Japanese NAFLD patients for excluding advanced fibrosis. In our study, the AUROC for FIB-4 index for discrimination of stage 3 from stages 0–2 was not satisfied (0.7143) and was inferior to NAFIC score. As described above, the lower AUROC for FIB-4 index than in previous reports may be due to the fact that stage 4 patients were excluded from the present study, but were included in previous reports. As the number of subjects in the present study was small, we cannot conclude which parameters and scoring system is best for the estimation of severity of NAFLD, nor we can definitively state that npRQ is the best method for differentiation of NASH from NAFLD. However, Table 4 shows that npRQ, HOMA-IR, AUC glucose, fasting glucose, insulin and NAFIC scores all could provide useful information for the detection of NASH, including patients in the early fibrosis stage, whereas NAFLD fibrosis score and type IV collagen were useful for identification of advanced fibrosis. It is important to select these parameters and scoring systems according to the purpose: to differentiate early-stage NASH from NAFLD, or to detect advanced fibrosis.
Although we believed that npRQ is useful for the estimation of disease severity in NAFLD patients, unfortunately, npRQ measurement is not always possible in clinical practice. This is because indirect calorimetry for measurement of npRQ was primarily used in inpatient and research settings and not all clinicians are familiar with the equipment used. In addition, patients must be tested in the early morning after fasting, and 24-h urine specimens must be collected for calculation of npRQ. Okumura et al. investigated the use of serum biochemistry to predict npRQ in patients with viral cirrhosis. That report concluded that serum NEFA can be used to predict npRQ. However, in the present study, npRQ was not correlated with NEFA. This may be because the subjects in the present study were NAFLD patients without cirrhosis, and not viral cirrhosis patients with decreased glycogen storage. In NAFLD patients with glucose intolerance, HOMA-IR may be useful for prediction of npRQ without calorimetry, because these two parameters were negatively correlated.
In conclusion, npRQ is useful for the estimation of disease severity in NAFLD patients with glucose intolerance. It enables the detection of NASH with relatively early-stage fibrosis among NAFLD patients. As this measurement can provide useful information without the burden of blood collection, it should be included in clinical practice in addition to anthropometry and blood analysis. When an NAFLD patient exhibits low npRQ, the patient should undergo further examination such as a liver biopsy to allow for early intervention.