Significant correlations between severe fatty liver and risk factors for metabolic syndrome
Abstract
Background and Aim: It is known that ultrasonography (US) cannot differentiate between non-alcoholic fatty liver disease (NAFLD) and steatohepatitis. However, US can accurately estimate the severity of the steatosis. The clinical significance of severe hepatic fatty change by US has not been explored. The aim of this study was to investigate the relationship between the severity of the fatty liver, classified by US, and the degree of metabolic disorders with insulin resistance.
Methods: In 16 486 Taiwanese patients, severity of fatty change on US was classified as follows: group A (n = 6950), absence of fatty change; group B (n = 8694), mild; and group C (n = 842), severe fatty liver change. Biometabolic parameters included body mass index (BMI), blood pressure (BP), fasting plasma glucose, triglycerides, cholesterol, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and serum creatinine. Nominal logistic regression analysis was used to estimate the odds ratio for different degrees of fatty liver.
Results: The frequencies of obesity, hypertension, glucose intolerance and hypertriglyceridemia were all significantly higher in group C than in group A or B (P < 0.0001), and the mean values of BMI, BP, fasting glucose, triglyceride and ALT were also higher in group C (P < 0.0001). High BMI (≥30 kg/m2) appears to be the most important factor for progression from mild to severe fatty liver in both sexes.
Conclusions: The presence of severe fatty liver by US correlated significantly with the prevalence and degree of hypertension, abnormal glucose and triglyceride metabolism. Patients with severe fatty liver could be at an increased risk of atherosclerotic cardiovascular disease and should be screened regularly for metabolic disorders. The physician may also evaluate ALT and hepatic fat content by US in patients with metabolic syndrome. Evaluating the severity of fatty liver by US may be useful because it correlates with the status of hyperinsulinemia, the risks of developing cardiovascular disease, and the threshold for oxidative stress.
Introduction
Non-alcoholic fatty liver disease (NAFLD) is now recognized as one of the most common liver disorders in the USA, with a prevalence of 10–24%. A similar trend is also found in Europe and Japan.1,2 Defined as a fat accumulation in the liver exceeding 5-10% by weight in the absence of alcohol abuse, contributing medications and viral hepatitis,3,4 NAFLD can be detected as bright liver on ultrasonography (US) in the absence of viral hepatitis in a non-drinker. Its histological abnormality can vary from simple steatosis to steatohepatitis (NASH), to fatty infiltration with ballooning degeneration, fibrosis and cirrhosis. Approximately 24–30% of patients with fibrosing steatohepatitis may progress to cirrhosis and liver-related death.1,4–7
NAFLD has been recognized as the manifestation of insulin resistance/hyperinsulinemia related metabolic syndrome affecting the liver,8 and US has proved to be a sensitive, accurate and convenient diagnostic tool in detecting steatosis: its sensitivity ranges from 60-94% and its specificity from 84-95%.9 When the hepatic steatosis reaches 33%, the detection sensitivity is nearly 100%.9–11 Sonographic rankings have been documented to be compatible with the severity of histological hepatic steatosis.11–13
Day et al.14 proposed the ‘two hits’ theory, now widely accepted, to explain the progression from simple steatosis to NASH. The first ‘hit’ is peripheral insulin resistance, resulting in the steatotic liver. The presence of increased triglycerides renders the steatotic or fatty liver more sensitive to oxidative stress (the ‘second hit’). Any event that increases oxidative stress in the liver (such as iron overload,15 mitochondrial abnormality,16 cytochrome P450 2E1 overactivity,17,18 NFκ B activation,19 hypoadiponectinemia20,21) may potentially trigger a transition from simple steatosis to NASH. Thus, the level of triglycerides in the liver affects the threshold to the ‘second hit. Although US cannot differentiate between NAFLD and NASH,16 it is a reliable tool in determining the extent of fatty liver infiltration.11–13 In this study, we investigated the correlation between severity of fatty liver by US and biometabolic parameters. Furthermore, we strove to determine the most important factor (body mass index [BMI], triglycerides, cholesterol, fasting plasma glucose) contributing to the progression to severe sonographic fatty liver to enable the practitioner to initiate appropriate specific intervention.
Methods
In 2003, 19 145 Taiwanese adults, residents from five counties in South Taiwan, were evaluated for inclusion in this study at the Center of Health Examination, Kaohsiung Medical University Hospital. They underwent a detailed medical history (including medication and alcohol use), physical examination, laboratory assessment, and abdominal B-mode US (3.5 MHz convex transducer, Toshiba SSA-250; Toshiba, Tokyo, Japan) carried out by hepatologists trained at the same institution to ensure interobserver consistency.
Excluded were individuals under the age of 20 years, those with known alcohol abuse, those taking medications with tamoxifen or amiodarone, pregnant women, individuals with positive seromarkers for hepatitis B or C, and those with an aspartate aminotransferase (AST)/alanine aminotransferease (ALT) ratio ≥2. The final study population comprised 16 486 subjects (14 333 men [86.9%] and 2153 women) aged 20-89 years.
The diagnosis of fatty liver was based on the brightness of the liver on US in comparison with the kidney, vascular blurring of the hepatic vein trunk, and deep attenuation in the right hepatic lobe.9,22 The severity of fatty liver change was classified according to standardized ultrasonographic criteria:10 group A, normal liver, a normal echo texture and absence of fatty change; group B, mild fatty liver change, a mild increase in fine echoes in the parenchyma with slightly impaired visualization of intrahepatic vessels and diaphragm; group C, severe fatty liver change, marked increase in fine echoes in the parenchyma with poor or non-visualization of the intrahepatic vessel borders, diaphragm and posterior right lobe of the liver. Criteria to claim ‘severe’ fatty liver change are obvious and unambiguous.
Using a multichannel autoanalyzer, we measured serum levels of aminotransferase, alkaline phosphatase (ALK-P), gamma-glutamyl transpeptidase (GGT), cholesterol, triglyceride and fasting glucose. Impaired fasting glucose (IFG) was defined by the American Diabetes Association criteria revised in 2003 as levels of 100–125 mg/dL.23 Metabolic abnormalities were diagnosed following the Third Report of the National Cholesterol Education Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults, modified by the criteria of obesity proposed for Asians by the Steering Committee of the Regional Office for the Western Pacific Region of WHO (WPRO).24–26 In accordance with the WPRO, the Asian standards of BMI (kg/m2) are: underweight <18.5; normal 18.5–22.9; overweight 23–24.9; obese stage I 25–29.9; and obese stage II ≥ 30.
Data were analyzed by SAS 8.2 software (SAS Institute, Cary, NC, USA). Results of continuous variables were expressed as mean ± standard deviation. During analysis, gender, age and BMI were stratified by criteria proposed by WPRO. Among the three groups, gender distribution was compared by the χ2-test and metabolic variables by one-way ANOVA (fasting glucose, BMI, systolic blood pressure, diastolic blood pressure, serum cholesterol, serum triglyceride, serum creatinine, AST, ALT, ALK-P and GGT). Using nominal logistic regression, we investigated the association between different degrees of fatty liver change by US (groups A, B and C) and metabolic parameters (age, fasting glucose, BMI, cholesterol, triglycerides). Odds ratios of developing mild fatty liver change (progressing from group A to B) and severe fatty liver change (from group A to C) were computed by stratified risk levels and adjusted by age and gender. Statistical significance was set at P < 0.05 in Tables 1–3 and <0.0036 in Tables 4 and 5 (Bonferroni correction).
| Absent FL(n = 6950) | Mild FL(n = 8694) | Severe FL(n = 842) | P-value | |
|---|---|---|---|---|
| Age (years) | 44.5 ± 8.9 | 45.7 ± 8.0 | 45.4 ± 8.8 | <0.001 |
| BMI (kg/m2) | 22.9 ± 2.7 | 25.4 ± 2.8 | 28.8 ± 3.5 | <0.001 |
| Systolic BP (mmHg) | 123.2 ± 16.1 | 127.7 ± 15.8 | 134.2 ± 17.3 | <0.001 |
| Diastolic BP (mmHg) | 79.3 ± 10.9 | 82.8 ± 10.8 | 87.4 ± 10.9 | <0.001 |
| Fasting glucose (mg/dL) | 97.6 ± 24.5 | 103.2 ± 28.1 | 113.5 ± 37.1 | <0.001 |
| Cholesterol (mg/dL) | 183.7 ± 29.2 | 191.8 ± 30.5 | 198.5 ± 34.2 | <0.001 |
| Triglyceride (mg/dL) | 109.8 ± 68.5 | 165.3 ± 115.1 | 229.7 ± 171.5 | <0.001 |
| Serum creatinine (mg/dL) | 1.1 ± 0.45 | 1.1 ± 0.31 | 1.11 ± 0.2 | NS |
| AST (IU/L) (0–40) | 22.7 ± 11.2 | 26.1 ± 13.6 | 35.7 ± 17.6 | <0.001 |
| ALT (IU/L) (0–40) | 22.0 ± 17.4 | 30.4 ± 21.0 | 50.1 ± 27.7 | <0.001 |
| ALK-P (IU/L) (50–180) | 118.2 ± 33.7 | 122.3 ± 33.8 | 130.8 ± 36.1 | <0.001 |
| GGT (IU/L) (4–50) | 22.1 ± 27.0 | 30.0 ± 37.2 | 41.4 ± 35.3 | <0.001 |
- Normal reference is expressed in parenthesis. Statistical analysis used one way ANOVA. ALT, alanine aminotransferase; ALK-P, alkaline phosphatase-P; AST, aspartate aminotransferase; BMI, body mass index; FL, fatty liver; GGT, gamma glutamyl transpeptidase; NS, not significant.
| Absent FL(n = 6645) | Mild FL(n = 8377) | Severe FL(n = 809) | P-value | |
|---|---|---|---|---|
| IFG | 2012 (30.3%) | 3312 (39.5%) | 377 (46.6%) | <0.001 |
| DM | 231 (3.5%) | 635 (7.6%) | 139 (17.2%) | <0.001 |
| HyperTG | 1171 (16.8%) | 3739 (43.0%) | 558 (66.3%) | <0.001 |
| H/T (≥130/85 mmHg) | 2918 (41.9%) | 4729 (54.4%) | 603 (71.6%) | <0.001 |
| BMI ≥25 kg/m2 | 1467 (22.1%) | 4629 (55.3%) | 729 (90.1%) | <0.001 |
| AST or ALT ≥1 × ULN | 493 (7.1%) | 1877 (21.6%) | 506 (60.1%) | <0.001 |
- Statistical analysis was calculated by the χ2-test. AST, aspartate aminotransferase; ALT, alanine aminotransferase; BMI, body mass index; DM, diabetes mellitus; FL, fatty liver; H/T, hypertension; HyperTG, hypertriglyceridemia with serum triglyceride ≥150 mg/dL; IFG, impaired fasting glucose; ULN, upper limit of normal reference.
| BMI (kg/m2) Population (%) | <18.5 (2.3%) | 18.5–22.9 (28.7%) | 23–24.9 (27.1%) | 25–29.9 (37.0%) | ≥30 (4.8%) | P-value |
|---|---|---|---|---|---|---|
| NAFLD | 38 (10%) | 1503 (32.1%) | 2546 (57.6%) | 4637 (76.7%) | 721 (92.2%) | <0.0001 |
| IFG | 78 (21.1%) | 1411 (30.3%) | 1574 (35.6%) | 2350 (37.9%) | 353 (44.0%) | <0.0001 |
| DM | 7 (1.9%) | 160 (3.4%) | 226 (5.1%) | 508 (8.2%) | 113 (14.1%) | <0.0001 |
| HyperTG | 14 (3.8%) | 746 (16.0%) | 1444 (32.6%) | 2805 (45.2%) | 478 (59.6%) | <0.0001 |
| H/T | 68 (18.4%) | 1553 (33.4%) | 2758 (62.3%) | 3881 (62.6%) | 621 (77.4%) | <0.0001 |
- Shows a significantly and positively linear correlation of BMI concomitant with NAFLD, IFG, diabetes, hypertriglyceridema and hypertension (P < 0.0001 by χ2-test with linear trend). DM, diabetes mellitus; H/T, hypertension; HyperTG, hypertriglyceridemia with serum triglyceride ≥150 mg/dL; IFG, impaired fasting glucose; NAFLD, non-alcoholic fatty liver disease.
| Mild FL vs absent FL | Severe FL vs absent FL | |||||
|---|---|---|---|---|---|---|
| OR | (95%CI) | P-value | OR | (95% CI) | P-value | |
| Fasting glucose <100 | 1.0 | 1.0 | ||||
| 100 ≤ fasting glucose <126 | 1.4 | (1.3–1.5) | <0.0001 | 2.0 | (1.7–2.4) | <0.0001 |
| Fasting glucose ≥126 | 1.7 | (1.4–2.0) | <0.0001 | 3.4 | (2.6–4.6) | <0.0001 |
| BMI <22.9 | 1.0 | 1.0 | ||||
| 23 ≤ BMI <24.9 | 2.7 | (2.4–2.9) | <0.0001 | 12.4 | (6.4–27.9) | <0.0001 |
| 25 ≤ BMI <29.9 | 5.7 | (5.1–6.2) | <0.0001 | 69.8 | (37.0–154.3) | <0.0001 |
| BMI ≥30 | 12.9 | (9.6–17.6) | <0.0001 | 805.6 | (399.5–1866.6) | <0.0001 |
| Cholesterol <200 | 1.0 | 1.0 | ||||
| 200 ≤ cholesterol <239 | 1.1 | (1.1–1.3) | 0.0013 | 1.2 | (1.0–1.5) | 0.0204 |
| Cholesterol ≥240 | 1.3 | (1.1–1.7) | 0.0046 | 1.8 | (1.3–2.5) | 0.0013 |
| Triglyceride <150 | 1.0 | 1.0 | ||||
| 150 ≤ triglyceride <200 | 2.2 | (2.0–2.4) | <0.0001 | 3.1 | (2.5–3.9) | <0.0001 |
| 200 ≤ triglyceride <400 | 3.1 | (2.7–3.5) | <0.0001 | 5.6 | (4.5–6.9) | <0.0001 |
| Triglyceride ≥400 | 3.6 | (2.7–5.0) | <0.0001 | 9.2 | (6.1–14.0) | <0.0001 |
- Statistical analysis used nominal logistic regression. Unit was mg/dL for plasma glucose, kg/m2 for body mass index (BMI), mg/dL for serum cholesterol and triglyceride. FL, fatty liver.
| Mild FL vs absent FL | Severe FL vs absent FL | |||||
|---|---|---|---|---|---|---|
| OR | (95%CI) | P-value | OR | (95%CI) | P-value | |
| Fasting glucose <100 | 1.0 | 1.0 | ||||
| 100 ≤ fasting glucose <126 | 1.6 | (1.3–2.0) | <0.0001 | 2.0 | (0.9–4.7) | 0.1159 |
| Fasting glucose ≥126 | 3.0 | (1.4–6.9) | 0.0049 | 4.3 | (1.1–17.0) | 0.0379 |
| BMI <22.9 | 1.0 | 1.0 | ||||
| 23 ≤ BMI <24.9 | 2.8 | (2.2–3.5) | <0.0001 | 11.6 | (2.6–80.0) | 0.0031 |
| 25 ≤ BMI <29.9 | 5.3 | (3.8–7.5) | <0.0001 | 77.8 | (20.7–505.9) | <0.0001 |
| BMI ≥ 30 | 12.3 | (5.2–36.3) | <0.0001 | 626.4 | (123.2–5028.2) | <0.0001 |
| Cholesterol <200 | 1.0 | 1.0 | ||||
| 200 ≤ cholesterol <239 | 1.2 | (1.0–1.5) | 0.1151 | 1.1 | (0.5–2.5) | 0.7886 |
| Cholesterol ≥240 | 0.8 | (0.5–1.5) | 0.5551 | 3.6 | (1.1–11.3) | 0.0349 |
| Triglyceride <150 | 1.0 | 1.0 | ||||
| 150 ≤τtriglyceride <200 | 2.6 | (1.7–4.0) | <0.0001 | 4.5 | (1.53–12.25) | 0.004 |
| 200 ≤ triglyceride <400 | 2.5 | (1.4–4.4) | 0.0019 | 14.1 | (5.3–37.5) | <0.0001 |
| Triglyceride ≥400 | 3.0 | (0.4–58.5) | 0.3358 | 12.6 | (0.4–430.0) | 0.1235 |
- Statistical analysis used nominal logistic regression. Unit was mg/dL for plasma glucose, kg/m2 for body mass index (BMI), mg/dL for serum cholesterol and triglyceride. FL, fatty liver.
Results
In our 16 486 subjects sonographic fatty liver change was demonstrated in 9536, with 52.7% of the total group classified as mild and 5.1% as severe. Table 1 summarizes comparisons of single variables among the three groups. For all but serum creatinine, correlations were significant, showing a positive parallel relationship between the degree of fatty infiltration and metabolic change.
The prevalence of IFG, diabetes, hypertriglyceridemia, hypertension and obesity also rose significantly as the degree of fatty infiltration increased from absent to severe. Group C subjects had a remarkably high prevalence of metabolic disorders (Table 2), and even group B subjects were notable for concomitant diabetes (7.6%) and IFG (39.5%).
An unusually high percentage (41.8%) of the study population was obese (37.0% obese I and 4.8% obese II) (Table 3). Not surprisingly, the clinical picture of subjects with a BMI ≥ 25 kg/m2 was complicated with high rates of metabolic abnormalities, such as glucose intolerance, hypertriglyceridemia and hypertension. The prevalence of fatty liver change by US increased in a stepwise fashion from normal BMI to obese II, demonstrating an extremely significant parallel correlation between higher BMI and higher grade of fatty infiltration (P < 0.0001 by χ2-test with linear trend).
With nominal logistic regression analysis, anthropometric parameters (BMI) and metabolic biomarkers (fasting glucose, cholesterol, triglyceride) were compared among groups A, B and C stratified by gender and age (Tables 4 and 5). The most important factor to correlate with the degree of fatty infiltration in both genders was BMI. Serum triglyceride level was second, and glucose intolerance (IFG and diabetes) third. Serum cholesterol level had less impact on this progression.
Discussion
In this study, the prevalence of fatty liver by US was unusually high (57.8%) when compared with the 10–30% rates published for the USA, Europe and Indonesia.27,28 NAFLD can be detected as bright liver on US in the absence of viral hepatitis in a non-drinker. Previous investigators have found different rates of NAFLD among races29 and ethnicity no doubt plays a role.
Anthropometric data regarding body fat distribution show that when Asian (mostly Chinese) and Caucasian populations are compared Asians have more body fat and more prominent central distribution for a given BMI.26,30,31 This may explain why our population has a higher prevalence of NAFLD than that reported for Caucasians.
In the ‘two hits’ theory proposed by Day et al.14 and now widely accepted,32 the first hit in the progression from simple steatosis to NASH is peripheral insulin resistance, which leads to increased lipolysis and delivery of free fatty acids (FFA) to the liver. These events may in turn result in triglyceride accumulation,33 detected as bright liver on US. This ‘second hit’ is required for the progression from steatosis to NASH.
The accumulation of triglycerides in hepatocytes will increase fatty acid β-oxidation. When this occurs in the presence of mitochondrial abnormalities (one example of the ‘second hit’), it would lead to free radical formation with consequent cell injury, inflammation and fibrosis.34 Without this ‘second hit’, the steatotic liver could remain stable and not progress to NASH.35 However, such a liver has an increased sensitivity to oxidative stress and any condition that increases it could act as a trigger. Diabetes mellitus, which is an independent factor for NASH, also increases the in vivo activity of cytochrome P450 2E1 in humans.36 These ‘second hits’ share a common character—augmented production of reactive oxygen species from enhanced oxidative stress and consequently increased hepatic triglyceride accumulation—and all have been associated with NASH.15–21,36–38
In Table 1, the presence of severe fatty liver change by US was significantly correlated with abnormal biometabolic parameters for metabolic syndrome. Thus, measuring the severity of hepatic triglyceride involvement may provide physicians with a sense of the threshold to oxidative stress. This information may help to identify patients whose livers are vulnerable to the ‘second hits’, and initiating a weight-loss program as soon as possible can be beneficial when hepatic necroinflammation and apoptosis are not yet prevalent.
Another question is which metabolic factor we should target to halt the progression to severe steatotic liver. Is hypertriglyceridemia, hypercholesterolemia, impaired fasting glucose or increased BMI the most relevant? As tumor necrosis factor-alpha (TNF-α) is elevated in the obese,37 and insulin resistance and beta-cell dysfunction can result,4 this can impair fasting glucose.
In our study, obesity was found to be the most important metabolic factor related to the progression of sonographic fatty liver severity. A direct correlation between the severity of fatty liver and BMI was clearly demonstrated for both genders (Tables 4 and 5). For example, a male subject with a BMI characterized as obese I (25–29.9 kg/m2) had a 69.8 times greater chance of developing severe fatty liver than a male with a normal BMI (see Table 4), and the odds soared to 805.6 times when the BMI was obese level II (>30 kg/m2). These results support the observation of Adams et al. that detection of NAFLD with intervention to halt progression should begin earlier, especially in people with pre-diabetes.39
Histopathological findings remain the gold standard for diagnosis and prognosis of NAFLD. However, they cannot be applied to a large-scale population.1,12,15 Abdominal US is non-invasive and easily undertaken in the clinical setting. Although US cannot differentiate between NAFLD and NASH,16 it does depict the severity of the steatosis.16,39–41 Because most presentations of NASH consist of subtle inflammatory reactions without major changes in liver morphology,16 it is not unexpected that US will not distinguish between it and steatosis. However, US is believed, at least by many scholars, to be a reliable tool in determining the extent of fatty liver infiltration.11–13 The sensitivity ranges from 60-94% and the specificity from 84-95%.9 If the hepatic steatosis reaches 33%, sensitivity approaches 100%.9–11 Ultrasonographic rankings have been shown to be compatible with the severity of histological hepatic steatosis.11–13 One sophisticated study demonstrated that US attenuation depends mainly on fatty infiltration of the liver, and to a lesser extent on fibrosis, but not on water content.40
We suggest that the obese, regardless of age, should undergo abdominal US. A finding of mild fatty liver by US indicating more than 30% steatotic hepatocytes may not pose an immediate threat, but it should prompt the clinician to counsel the patient that the liver is already overloaded with triglycerides and to outline the likely consequences should overweight persist. In addition, even in patients with normal BMI, if severe fatty liver change is found on US, the metabolic biomarkers should be evaluated to identify pre-diabetics. In Table 3, 10 to 32% of subjects with normal BMI (<23 kg/m2) developed fatty liver change by US.
Convincing the obese patient to begin a weight-loss regimen in the absence of an imminent health crisis can be supported by the US evidence and the clinician's explanation of why this should be considered sooner rather than later. As yet, no pharmacotherapy has proved effective for late-stage NAFLD (NASH and cirrhosis). Of three randomized trials42–44 (with antioxidants [vitamins E and C], ursodeoxycholic acid, metformin), all failed to alleviate the histological change. Additionally, metabolic disorders usually worsen with advanced NAFLD:11 the term ‘hepatogenous diabetes’ is used to describe this condition. In our population the high prevalence of steatotic liver coexisting with IFG (40.2%; see Table 2) suggests that many adult Taiwanese with NAFLD will develop hepatogenous diabetes in the future. NAFLD has been associated with insulin resistance/hyperinsulinemia, and insulin can stimulate the proliferation of arteriolar smooth muscle cells. Thus NAFLD patients might be at an increased risk of atherosclerotic cardiovascular disease. From this study, we demonstrated that the severity of the fatty infiltration correlates very well to the degree of the metabolic disorders with insulin resistance. Evaluating the severity of fatty liver may provide the physician with a piece of useful information about patients' hyperinsulinemia status, the risks of developing cardiovascular disease and the threshold to oxidative stress.
The present study does have some limitations. First, we tried our best to exclude subjects with a history of habitual alcohol consumption and those with an AST/ALT ratio ≥2 in this study. However, the exact amount of alcohol consumed and its alcoholic content were difficult to assess. Also, fatty liver was defined as a fat accumulation in the liver exceeding 5-10% by weight, and when the hepatic steatosis reaches 33%, the detection sensitivity by US is nearly 100%.9–11 Thus, the 80% sensitivity in average of US for detection of steatotic liver may have resulted in some false-negative readings.
In conclusion, our data demonstrate that the presence of severe fatty liver by US is significantly associated with worsened biometabolic parameters. Measuring the severity of fatty infiltration in the liver may provide physicians with a sense of the threshold to the ‘second hits’ for NAFLD, that is, those patients whose livers are vulnerable to oxidative stress. Obesity, especially morbid obesity (obese II), was the strongest factor associated with the transition from normal liver parenchyma to severe steatotic liver. We believe weight loss before the onset of the ‘second hit’ is beneficial, as irreversible cell loss has not yet occurred.
Acknowledgments
We are indebted to Mr Horng-Ming Lu for his assistance with the computer calculations. We thank Horng-Ming Lu, who is a candidate for the master degree of Graduate Institute of Pharmaceutical Science, for his capable assistance in computer calculations.
