SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Background  Cross-sectional studies suggest insulin resistance is strongly associated with hepatic steatosis and fibrosis in patients with chronic hepatitis C (CHC), which might affect the efficacy of antiviral therapy.

Aim  To investigate retrospectively the impact of insulin resistance on treatment response in Chinese genotype 1 CHC patients receiving a 24-week course therapy with peginterferon α-2b/ribavirin.

Methods  A total of 133 biopsy-proven CHC patients were enrolled for analyses. Insulin resistance was evaluated by homeostasis model assessment of insulin resistance (HOMA-IR). Hepatic fibrosis was graded by the METAVIR scoring system.

Results  Mean HOMA-IR progressively elevated along with the severity of hepatic fibrosis (F1–F2 fibrosis: 2.55 ± 0.16 vs. F3–F4 fibrosis: 3.61 ± 0.20, < 0.001). Compared with patients with sustained virological response (SVR), patients without SVR had significantly higher percentages of F3–F4 fibrosis (62.2% vs. 21.6%, < 0.001) and baseline high viral load (≥600 000 IU/mL; 64.4% vs. 35.6%, = 0.038). In addition, patients without SVR had significantly higher plasma levels of insulin (15.03 ± 0.89 vs. 10.19 ± 0.55 μU/mL, < 0.001) and HOMA-IR values (3.76 ± 0.23 vs. 2.50 ± 0.15, < 0.001). Multivariate analyses showed that F1–F2 fibrosis (odds ratio: 4.49, = 0.001), HOMA-IR < 2 (odds ratio: 7.15, = 0.005) and pre-treatment hepatitis C virus RNA < 600 000 IU/mL (odds ratio: 3.26, = 0.012) were the independent factors associated with SVR.

Conclusions  Insulin resistance is a major determinant of SVR in genotype 1 CHC patients receiving peginterferon α-2b/ribavirin. Strategies to modify insulin resistance may be effective in enhancing SVR before or during anti-viral therapy.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Emerging evidence reveals that type-II diabetes mellitus occurs more frequently in patients with chronic hepatitis C virus (HCV) infection than other forms of chronic inflammatory liver diseases.1–4 In addition, insulin resistance, the reliable predictor of type-II diabetes,5, 6 has been found in part in nondiabetic chronic hepatitis C (CHC) patients and can occur in the early stage of infection.7–9 To date, the mechanisms responsible for insulin resistance in patients with chronic HCV infection are not fully understood. However, recently published data suggested that it might be attributed by a multifactorial phenomenon, which includes impaired hepatic insulin clearance in advanced liver fibrosis,10–12 induced by inflammatory cytokines which are incited by HCV infection,13, 14 and directly caused by HCV proteins.15–18

Several baseline host or viral characteristics have been found to correlate with sustained response in CHC patients receiving therapy with interferon (IFN)/pegylated IFN or in combination with ribavirin. These factors include younger age, female gender, lesser hepatic steatosis, HCV genotype non-1, lower baseline viral levels, less fibrosis or inflammation on liver biopsy and lower body weight/body mass index (BMI).19–24 Cross-sectional studies suggest that insulin resistance is strongly associated with obesity,8 hepatic steatosis9, 25 and more advanced liver fibrosis,7–9 all of these constituting negative predictors of treatment response. Some recently published studies suggested that insulin resistance might impair the sustained response rate to pegylated IFN plus ribavrin in CHC patients.26–28 However, none of the above studies focused on the same genotype of CHC patients and used the same kind of pegylated IFN. Therefore, we retrospectively analysed the impact of insulin resistance on treatment response in genotype 1 Chinese CHC patients treated with peginterferon α-2b plus ribavirin.

Patients and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Eligible patients

Between November 2004 and December 2006, consecutive patients with genotype 1 CHC infection who received therapy with pegylated IFN α-2b plus ribavirin at the Liver Clinic of Taipei Veterans General Hospital were eligible candidates for this study. CHC was defined as follows: (i) patients with positive serum antibody to HCV and detectable serum HCV RNA; (ii) patients having undergone a liver biopsy within 1 year previous to entry into the study consistent with chronic hepatitis; and (iii) patients with elevated serum alanine aminotransferase (ALT) defined as ≥2 × (upper limit of normal) for at least two measurements within 6 months preceding entry into the study. Patients were excluded if they had positive hepatitis B surface antigen, previous liver transplantation, established a diagnosis of diabetes now taking anti-diabetic medicine or fasting sugar >125 mg/dL, human immunodeficiency virus infection, low blood cell counts at baseline, autoimmune diseases or other contraindications for IFN and ribavirin therapy. Patients who early terminated anti-viral therapy were excluded for analyses. This study was performed in accordance with the principles of Good Clinical Practice, the principles of the Declaration of Helsinki and its appendices and local and national laws.

Treatment regimen and dose modifications

Patients were treated by peginterferon α-2b (PEG-Intron; Schering-Plough Inc., Kenilworth, NJ, USA) at a dose of 1.5 μg/kg/week subcutaneously plus oral ribavirin 800–1400 mg/day (Rebetol; Schering-Plough Inc.). The dose of ribavirin was based on body weight (800 mg for weight ≤50 kg, 1000 mg for weight ≤65 kg, 1200 for weight ≤80 kg and 1400 mg for weight >80 kg) and was orally administered with food in two divided doses per day. Subjects were treated for 24 weeks and followed for another 24 weeks after the end of treatment. The Bureau of National Health Insurance reimbursed the cost of peginterferon and ribavirin.

During treatment, patients were assessed as out-patients at week 0, 2, 4, 6, 8, then every 4 weeks for the duration of treatment and also at 4, 12 and 24 weeks after the end of therapy. At each visit, physical examination was performed and adverse events were recorded. Biochemical and haematological testing was performed by commercial assays. Serum HCV RNA at baseline was measured by commercially available quantitative assay (Versant HCV RNA 3.0 Quantitative Assay; Bayer Corporation, Tarrytown, NY, USA) with a sensitivity of 615 IU/mL (∼3200 copies/mL).29 Serum HCV RNA at the end-of-treatment and at 24 weeks after therapy was measured by qualitative assay (Cobas Amplicor HCV v2.0 Assay; Roche Diagnostics, Branchburg, NJ, USA) with the detection limit of 50 IU/mL to determine treatment response. HCV genotyping was performed as previously described.30 Degree of liver fibrosis was graded by the METAVIR scoring system.31 Grades of necroinflammation was determined by the modified HAI score: minimal–mild (score 1–8), moderate/severe (score 9–18).32 Steatosis was assessed as the percentage of hepatocytes containing macrovesicular fat droplets. It was graded as follows: 0 = no steatosis, 1 = mild: 1–20% of hepatocytes affected; 2 = moderate: 21–40% of hepatocytes affected; 3 = severe: >40% of hepatocytes affected.25

The dose of peginterferon was decreased by 25% and the dose of ribavirin was lowered to 600 mg/day when severe adverse events occurred or when laboratory results showed haemoglobin <9 g/dL in subjects with no cardiac disease, haemoglobin decrease of >2 g/dL in subjects with cardiac disease, white cell count (WBC) <1500/mm3 (neutrophil <750/mm3) or platelet count <50 000/mm3. Full doses could be resumed when the event was resolved. If the event persisted, both drugs were discontinued. Therapy was permanently discontinued for life-threatening events or when laboratory results showed haemoglobin <8 g/dL in subjects with no cardiac disease, haemoglobin <12 g/dL in subjects with cardiac disease after 4 weeks of dose reduction, WBC <1000/mm3 (neutrophil <500/mm3), platelet count <30 000/mm3 or serum creatinine >2.0 mg/dL.

Assessment of treatment outcome

End-of-treatment virological response was defined as disappearance of serum HCV RNA (<50 IU/mL) at the end of treatment. Sustained virological response (SVR) was defined as this phenomenon persisting to 24 weeks after the end of treatment.

Laboratory investigations for insulin resistance

For each patient, insulin resistance was evaluated before the initiation of anti-viral therapy. An overnight fasting blood sample was drawn to determine the plasma levels of insulin by commercialized two-site sandwich immunoassay using direct chemiluminescent technology (ADVIA Centaur Insulin Assay; Bayer Diagnostics, Leverkusen, Germany). The lower detection limit of this assay was 0.1 μU/mL. Insulin resistance was determined by the homeostasis model assessment (HOMA) method using the following equation: insulin resistance (HOMA-IR) = fasting insulin (μU/mL) × fasting glucose (mmol/L)/22.5.33 HOMA-IR has been validated in comparison with the euglycemic/hyperinsulinemic clamp technique in both diabetic and nondiabetic patients.34

Statistical analyses

Results were expressed as mean ± S.E. Data were analysed using the spss, version 10.0 software package (SPSS Inc., Chicago, IL, USA). Statistical analyses were performed using Chi-square and Fisher’s exact tests for categorical variables. Independent t-tests were used for continuous variables. Multivariate analyses with stepwise logistic regression were used to determine the independent factors that correlate with treatment response. Results were considered statistically significant at < 0.05.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Baseline characteristics of enrolled patients

During the enrollment period, a total of 163 genotype 1 CHC patients started therapy with peginterferon α-2b plus ribavirin at our Clinic. Seventeen patients with the diagnosis of diabetes were excluded and a total of 13 patients early terminated therapy, therefore, a total of 133 patients were eligible for analyses. The reasons for early termination were as follows: intolerance of side effects: four (23.1%), laboratory abnormalities including severe anaemia, thrombocytopenia and elevated total bilirubin levels: four (23.1%), lost to follow-up: two (15.4%), pyogenic infection: one (7.7%), gastrointestinal bleeding: one (7.7%), and thyrotoxicosis: one (7.7%). Regarding the baseline characteristics, 70 (52.6%) of them were male with a mean age of 53.4 ± 1.1 years old (range: 22–78). Almost all of them (94.7%) were infected by subtype 1b and the mean pre-treatment HCV RNA level was 5.66 ± 0.09 log10 IU/mL. Liver histology at baseline showed 86 (64.7%) of them with F1–F2 fibrosis and the percentage to have liver cirrhosis was 16.5%. The baseline information of enrolled patients regarding liver biochemistries, haematology and virological tests are summarized in Table 1.

Table 1.   Baseline characteristics of enrolled patients
CharacteristicsResults (= 133)
  1. ALT, alanine aminotransferase; AST, aspartate aminotransferase; HCV, hepatitis C virus; M, million.

Mean age (years)53.4 ± 1.1
Male/female (%) 70/63 (52.6/47.4)
Mean height (cm)163.1 ± 0.8
Mean weight (kg)63.9 ± 1.0
Mean body mass index (kg/m2) 23.89 ± 0.28
Mean ALT (IU/L) 143 ± 7
Mean AST (IU/L) 100 ± 6
Mean total bilirubin (mg/dL) 0.79 ± 0.03
Mean fasting sugar (mg/dL) 99.2 ± 0.9
Mean white cell count (/cumm)5492 ± 130
Mean Hgb (g/dL) 13.95 ± 0.14
Mean platelets (×1000/cumm)181.7 ± 4.9
Mean insulin (μU/mL) 11.83 ± 0.51
Mean HOMA-IR2.93 ± 0.14
Subtypes of HCV (1a/1b)7/126 (5.3%/94.7%)
Mean HCV RNA (log10 IU/mL) 5.66 ± 0.09
Baseline HCV RNA ≥ 600 000 IU/mL/<600 000 IU/mL (%) 69/64 (51.9%/48.1%)
Inflammation grading: minimal–mild (score 1–8)/moderate–severe (score 9–18)51/82 (38.3%/61.7%)
Fibrosis grading: F1–F2/F3–F4 (%) 86/47 (64.7/35.3)
With cirrhosis (%) 22 (16.5)
Steatosis grading: absent–mild/moderate–severe (%)109/24 (82.0/18.0)
Mean initial ribavirin dose (mg/kg)16.77 ± 0.18

During the treatment period, a total of 21 (15.8%) patients experienced reduction dose of peginterferon and a total of 32 (24.1%) patients experienced dose reduction of ribavirin. Eleven (8.3%) patients failed to maintain the cumulative dose of both agents above 80% after finishing therapy.

Relationship between insulin resistance and degree of hepatic fibrosis (Figure 1)

image

Figure 1.  Relationship between insulin resistance and degree of hepatic fibrosis. The box plots show the 10th, 25th, 50th, 75th and 90th percentiles and outliers (circles). Mean HOMA-IR value is indicated as a dotted line.

Download figure to PowerPoint

The relationship between HOMA-IR value and hepatic fibrosis was illustrated in Figure 1. Mean HOMA-IR progressively elevated along with the severity of hepatic fibrosis with significant differences (F1–F2 fibrosis: 2.55 ± 0.16 vs. F3–F4 fibrosis: 3.61 ± 0.20, < 0.001). Histological findings showed that a higher percentage of moderate/severe fibrosis can be found in patients with F3–F4 fibrosis than in those with F1–F2 fibrosis (23.4% vs. 15.1%); however, it did not reach a significant level.

Comparisons of the clinical and virological characteristics between patients with or without SVR

After therapy, the overall SVR rate of the enrolled population was 66.2% (88/133). Twenty-nine (21.8%) patients relapsed after cessation of therapy and 16 (12.0%) patients belonged to nonresponders. Table 2 compares the clinical and virological characteristics between patients with or without SVR. Mean age, gender and baseline ALT levels were comparable between these two groups. Compared to patients with SVR, patients who did not achieve SVR had a higher BMI (= 0.006), aspartate aminotransferase (AST) value (= 0.017), total bilirubin level (= 0.040) and lower platelet count (< 0.001). The mean fasting glucose was comparable between these two groups. However, patients without SVR had significantly higher plasma levels of insulin (15.03 ± 0.89 vs. 10.19 ± 0.55 μU/mL, < 0.001) and HOMA-IR values (3.76 ± 0.23 vs. 2.50 ± 0.15, < 0.001) than patients achieving SVR.

Table 2.   Comparison of the clinical and virological characteristics between patients who achieved SVR and who did not
 Patients with SVR (= 88)Patients without SVR (= 45)P-value
  1. ALT, alanine aminotransferase; AST, aspartate aminotransferase; SVR, sustained virological response.

Mean age (years)53.2 ± 1.353.8 ± 1.7NS
Male (%) 48/88 (54.5)22/45(48.9)NS
Mean body mass index (kg/m2) 23.40 ± 0.3824.84 ± 0.340.006
Mean ALT (IU/L) 137 ± 9155 ± 13NS
Mean AST (IU/L) 88 ± 6122 ± 130.017
Mean total bilirubin (mg/dL) 0.74 ± 0.040.88 ± 0.060.040
Mean fasting sugar (mg/dL) 98.2 ± 1.2101.1 ± 1.4NS
Mean insulin (μU/mL) 10.19 ± 0.5515.03 ± 0.89<0.001
Mean HOMA-IR2.50 ± 0.153.76 ± 0.23<0.001
Mean white cell count (/cumm)5609 ± 1605264 ± 224NS
Mean Hgb (g/dL) 13.9 ± 0.213.9 ± 0.2NS
Mean platelets (×1000/cumm)195.3 ± 5.6155.1 ± 7.9<0.001
HCV subtype 1a/1b (%)5/83 (5.7%/94.3)2/43 (4.4%/95.6)NS
Mean HCV RNA (log10 IU/mL) 5.53 ± 0.125.91 ± 0.14NS
Baseline HCV RNA ≥600 000 IU/mL/<600 000 IU/mL (%)40/48 (45.5%/54.5)29/16 (64.4%/35.6)0.038
Fibrosis grading: F1–F2/F3–F4 (%)69/19 (78.4/21.6)17/28 (37.8/62.2)<0.001
With cirrhosis (%)8/88 (9.1)14/45 (31.1)0.001
Steatosis grading: absent–mild/moderate–severe (%) 75/13 (85.2/14.8)34/11 (75.6/24.4)NS

Baseline liver histology was significantly associated with treatment response. In comparison with patients with SVR, a higher proportion of patients without SVR belonged to F3–F4 fibrosis (62.2% vs. 21.6%, < 0.001) and 31.1% of them with liver cirrhosis (= 0.001). A lower possibility to achieve SVR was found in patients with moderate/severe steatosis compared with absent/mild steatosis group (54.2% vs. 68.8%); however, it failed to reach a significant level. In addition, the percentages to have baseline high viral load (≥600 000 IU/mL; 64.4% vs. 35.6%, = 0.038) were markedly different between these two groups.

Relationship between insulin resistance and treatment response (Figure 2)

image

Figure 2.  Distribution of HOMA-IR value according to treatment response. The box plots show the 10th, 25th, 50th, 75th, and 90th percentiles and outliers (circles). Mean HOMA-IR value is indicated as a dotted line.

Download figure to PowerPoint

Figure 2 depicts the distribution of HOMA-IR value by box-plot according to treatment response. The mean HOMA-IR value of patients with SVR, relapser and nonresponder was 2.50 ± 0.15, 3.66 ± 0.28, and 3.95 ± 0.43, respectively. Significant differences were observed when SVR group was compared with relapser (= 0.002) and nonresponder groups (= 0.001).

Impact of insulin resistance on SVR in different categories of fibrosis (Figure 3)

image

Figure 3.  Impact of insulin resistance on sustained virological response in patients with different fibrosis categories.

Download figure to PowerPoint

In this study, a negative relationship was found between treatment response and insulin resistance. The overall SVR rates in patients with HOMA-IR < 2, 2 < HOMA-IR < 4 and HOMA-IR > 4 were 92.7%, 61.0% and 42.4%, respectively. This phenomenon was more prominent in patients with F1–F2 fibrosis than in patients with F3–F4 fibrosis (Figure 3). The SVR rate in F1–F2 fibrosis patients with HOMA-IR < 2, 2 ≤ HOMA-IR < 4, and HOMA-IR ≥ 4 were 94.6% (35/37), 75.0% (27/36) and 53.8% (7/13), respectively (= 0.004). In F3–F4 fibrosis category, SVR rate for patients with HOMA-IR < 2, 2 ≤ HOMA-IR < 4 and HOMA-IR ≥ 4 were 75.0% (3/4), 39.1% (9/23) and 35.0% (7/20), respectively.

Independent factors associated with SVR in Chinese genotype 1 CHC patients treated with pegylated IFN plus ribavirin

Potential baseline characteristics that may be associated with treatment response, including gender, age, BMI, HCV subtypes, pre-treatment HCV RNA levels, hepatic fibrosis, steatosis and degree of insulin resistance were analysed by multivariate logistic regression. Our results suggested that F1–F2 fibrosis (odds ratio: 4.49, = 0.001), HOMA-IR < 2 (odds ratio: 7.15, = 0.005) and pre-treatment HCV RNA< 600 000 IU/mL (odds ratio: 3.26, = 0.012) were the independent factors associated with SVR in Chinese genotype 1 CHC patients treated with pegylated IFN α-2b plus ribavirin (Table 3).

Table 3.   Independent factors associated with sustained virological response in Chinese genotype 1 chronic hepatitis C patients receiving pegylated interferon plus ribavirin
FactorsOdds ratio95% confidence intervalP-value
F1–F2 fibrosis4.491.81–11.150.001
HOMA-IR <27.151.84–27.860.005
Pre-treatment hepatitis C virus RNA <600 000 IU/mL3.261.29–8.200.012

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In this study, the SVR rates for Chinese genotype 1 CHC patients finished a 6-month course of therapy with peg-interferon α-2b and ribavirin was 66.2% by per-protocol analyses. These findings were similar to what had been previously published in a Taiwan multi-center trial,35 which guided the duration and dosing regimen of this study. As compared with previous large-scaled studies in Caucasians,23, 24, 36 our results in SVR were better, even with a shorter duration of therapy. The possible reasons explaining the above phenomenon are as follows: First, as confirmed by previous results,23, 36 the dose of ribavirin was a very important factor in determining treatment response. In comparison with previous Caucasian studies,23, 24, 36 the average dose of ribavirin per kg body weight was higher in our patients. In addition, our patients had lower mean BMI and obesity was confirmed as a negative factor for treatment response. Furthermore, it has been reported that compared with subtype 1a, CHC patients infected with subtype 1b had more rapid viral decline during pegylated IFN plus ribavirin therapy.37 As almost all our enrolled patients were infected by subtype 1b HCV, it is possible that favourable viral kinetics during combination therapy may play a role in better response. Another factor is probably ethnicity-related, as African-Americans responded poorer than Caucasians,38 although comprehensive, large-scale studies are needed to clarify this issue.

By homeostasis model assessment, our study confirmed that insulin resistance could be found in earlier stages of HCV infection. In concordance with previous reports,7–9 our study showed that HOMA-IR was positively associated with the degree of hepatic fibrosis. Growing data suggest that insulin resistance may aggravate hepatic fibrosis through several mechanisms. Hyperinsulinemia can directly stimulate hepatic stellate cells to proliferate and to secrete extracellular matrix.39 Furthermore, high glucose levels and hyperinsulinemia cause up-regulation of connective growth factor,40 a cytokine involved in the pathogenesis of fibrosing liver diseases.41 Therefore, a vicious cycle can develop after the occurrence of insulin resistance further aggravating the condition of hepatic fibrosis. In contrast, research studies that focused on the relationship between insulin resistance and hepatitis B infection are rare. Preliminary data suggested the phenomenon of insulin resistance was rare in chronic hepatitis B infection42 and might not be associated with fibrosis progression43 suggesting that a virus-specific pathogenesis might participate in the development of insulin resistance. However, more studies are needed to clarify this issue.

Compatible with previous large-scaled studies, our results supported that besides genotypes, the degree of hepatic fibrosis and baseline viral load are important factors associated with treatment response in CHC patients treated with IFN with or without combination with ribavirin. More importantly, by multivariate logistic regression, our data suggested that insulin resistance was a major determinant of SVR in genotype 1 CHC patients treated with a 24-week course of combination therapy. This finding was compatible with study results of Romero-Gomez et al.26 in which the treatment duration for genotype 1 patients was 48 weeks. According to the current study results, determination of the degree of insulin resistance along with other proven characteristics before initiation of anti-viral therapy can help us identify the difficult-to-treat population. Strategies to enhance medical adherence and early interventions with serotonin reuptake inhibitors or hematological growth factors to alleviate adverse effects may be helpful to improve the overall treatment response.

Our subgroup analyses showed that the influence of insulin resistance was more prominent in genotype 1 patients with less severe liver fibrosis. These results suggest that other than the correlation between insulin resistance and hepatic fibrosis, insulin resistance may affect IFN treatment response through other pathways. TNF-α, a cytokine that can be stimulated by CHC infection, has been shown to be involved in the pathogenesis of insulin resistance and response to IFN therapy. TNF-α has been shown to correlate with HOMA-IR values in CHC patients44, 45 and treatment with anti-TNF-α has been shown to restore insulin sensitivity in a model of HCV core transgenic mice.34 An in vitro study showed that TNF-α can stimulate the production of oxygen-free radicals,46 such as superoxide anion and hydroxyl radicals, subsequently leading to intracellular glutathion depletion.47 Depletion of cellular glutathion has been proposed as one possible mechanism responsible for HCV persistence.48 Human studies have also demonstrated that a higher plasma level of TNF-α correlates with poor response to IFN therapy.49, 50 In addition, insulin resistance was associated with increased hepatic expression of suppressor of cytokine signalling 3 (SOCS-3),37 a negative regulator of IFN-α signalling pathway.51 A recently published study confirmed that increased SOCS-3 expression correlated with treatment failure in genotype 1 CHC patients receiving anti-viral therapy.52 Further studies are warranted to investigate whether modification of insulin resistance can improve SVR in patients before or during therapy with pegylated IFN plus ribavirin.

In conclusion, our study found that the degree of hepatic fibrosis, insulin resistance and baseline HCV RNA level are the most important baseline characteristics associated with SVR in genotype 1 patients receiving peginterferon plus ribavirin therapy. Determination of HOMA-IR before anti-viral therapy can help us to predict SVR especially in patients with less severe liver fibrosis. Strategies to modify insulin resistance before or during combination therapy may be a feasible approach in enhancing the likelihood of treatment response, especially for genotype 1 patients.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The authors wish to thank Mei-Yi Chou for the excellent statistical assistance. Declaration of personal interests: C.-J. Chu has served as an invited speaker for Bristol-Myers Squibb, GlaxoSmithKline, Novartis, Roche Pharmaceuticals and Schering-Plough, and received research funding from GlaxoSmithKline and Schering-Plough. Declaration of funding interests: This study was supported in part by a grant from the National Science Council, Republic of China (NSC 94-2314-B-075-085), Taipei Veterans General Hospital (VGH 94-232, V96C1-114) and Szu-Yuan Research Foundation of Internal Medicine (96-008).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References