The number of patients who were diagnosed with chronic HCV infection between April 1975 and March 2006 in the Department of Hepatology, Toranomon Hospital, Tokyo, Japan was 11 500. Of these, 6150 patients met the following criteria: (1) age of 40-70 years; (2) features of chronic hepatitis or cirrhosis diagnosed by laparoscopy, liver biopsy, ultrasonography clinical features, and/or laboratory tests; (3) positive for anti-HCV antibody and HCV-RNA; (4) no history of treatment with antiviral agents; (5) negative for hepatitis B surface antigens (HBsAg), antinuclear antibodies, or antimitochondrial antibodies in serum, as determined by radioimmunoassay or spot hybridization; (6) no evidence of HCC nodules as shown by ultrasonography and/or computed tomography; (7) no underlying systemic disease, such as systemic lupus erythmatosus, rheumatic arthritis; (8) no cough or dyspnea after exercising; (9) no history of chronic lung disease. Patients with any of the following criteria were excluded from the study: (1) alpha-fetoprotein of 400 ng/mL or higher; (2) advanced and decompensated stage of cirrhosis with encephalopathy, icterus, or refractory ascites; (3) a short follow-up period of 6 mo or less; (4) development of HCC within 6 mo after the initiation of follow-up. These 6150 patients were regarded as in HCV-group. The 2050 patients that did not have the HCV marker and have the HBsAg marker were selected as a control and they comprised the hepatitis B virus (HBV)-group. The patients in HBV-group meet all of the above criteria but 3. Control patients were matched 1:3 with HCV positive patients for age and sex. We compared the differences of the cumulative development rate of IPF in both the HC-group and the HBV-group. Next, we assessed predictive factors for IPF in patients with hepatitis C. All of the studies were performed retrospectively by collecting and analyzing data from the patient records.

IPF was diagnosed by respiratory specialist based on the presence of at the least three of the following four diagnostic major criteria, as well as all of the following four minor criteria, as shown in Table 1. Diagnostic criteria of IPF were recommended by American Thoracic Society/European Respiratory Society[22]. We excluded hepatopulmonary syndrome by conventional chest radiographs or high-resolution computed tomography scans, electrocardiogram, and/or ultrasonic cardiography.

Anti-HCV was detected using a second-generation enzyme-linked immunosorbent assay (ELISA II) (Abbott Laboratories, North Chicago, IL). HCV-RNA was determined by the Amplicor method (Cobas Amplicor HCV Monitor Test, v2.0, Roche Molecular Systems, Inc., NJ). HBsAg was tested by radioimmunoassay (Abbott Laboratories, Detroit, MI). The used serum samples were stored -80°C at the first consultation. Diagnosis of HCV infection was based on detection of serum HCV antibody and positive RNA. The study started in April 1975. At that time, HCV testing was not available. Thus, the patients diagnosed as HCV-infected were tested after HCV testing became available.

Status of liver cirrhosis was mainly determined on the basis of peritoneoscopy and/or liver biopsy. Six thousand eight hundred and twenty six out of 8200 were diagnosed by peritoneoscopy and/or liver biopsy. Liver biopsy specimens were obtained using a modified Vim Silverman needle with an internal diameter of 2 mm (Tohoku University style, Kakinuma Factory, Tokyo, Japan), fixed in 10% formalin, and stained with hematoxylin-eosin, Masson’s trichrome, silver impregnation, and periodic acid-Schiff after diastase digestion. The size of specimens for examination was more than six portal areas. Baseline liver histology of chronic hepatitis was classified according to the extent of fibrosis, into four stages in progression order: stage 1, periportal expansion; stage 2, portoportal septa; stage 3, portocentral linkage or bridging fibrosis; stage 4, liver cirrhosis[23]. Remaining patients were diagnosed by clinical features, laboratory tests, and ultrasonographic findings. Ultrasonography was performed with a high-resolution, real-time scanner (model SSD-2000; Aloka Co., Ltd, Tokyo Japan. Logic 700 MR; GE-Yokokawa Medical Systems, Tokyo, Japan). The diagnosis of liver cirrhosis was defined as having a score of > 8 in ultrasonographical scoring system based on liver surface, liver parenchyma, hepatic vessel and spleen size as reported by Lin et al[24].

Patients were followed-up monthly to tri-monthly after the first medical examination in our hospital. Physical examination and biochemical tests were conducted at each examination together with regular check up using abdominal CT or US imaging in each patient. When a patient had any symptoms in relation to IPF (dry cough, dyspnea), we further explored the possibility of that patient having IPF. Three hundred thirty-four patients were lost to follow-up. Because the appearance of IPF and death was not identified in these 334 patients, they were considered as censored data in statistical analysis[25]. Moreover, patients treated with anti-viral agents were regarded as withdrawals at the time of starting antiviral agents.

Nonparametric procedures were employed for the analysis of background features of the patients, including the Mann-Whitney U test and χ² method. The cumulative appearance rate of IPF was calculated from the period of the first medical examination at our hospital to the appearance of IPF, using the Kaplan-Meier method. Differences in the development of IPF were tested using the log rank test. Independent factors associated with the incidence rate of IPF were analyzed by the Cox proportional hazard model. The following nine variables were analyzed for potential covariates for incidence of IPF at the time of first medical examination at our hospital: age, sex, state of liver disease (chronic hepatitis or liver cirrhosis), smoking index, total alcohol intake, platelet count, aspartate aminotransferase (AST), alanine aminotransferase (ALT), γ-glutamyl transpeptidase. P < 0.05 in two-tailed test was considered significant. Data analysis was performed using the SPSS computer program package (SPSS 11.5 for Windows, SPSS, Chicago, IL).

Table 2 shows the characteristics of the 8200 patients with HCV or HBV. There were no significant differences between the two groups with regard to sex ratio or age. However, there were significant differences in histopathological stage of the liver, AST, ALT, total bilirubin, total intake of alcohol, and smoking index. The 1594 (77.8%) out of 2050 in HBV-group and 5232 (85.1%) of 6150 in HCV-group were diagnosed by laparoscopy and/or liver biopsy. The breakdown of histological staging in HCV-group was as follows: stage 1, 2707; stage 2, 1188; stage 3, 300; stage 4 (liver cirrhosis), 1037. The outbreak of histological staging in HBV-group was as follows: stage 1, 705; stage 2, 439; stage 3, 157; stage 4 (liver cirrhosis), 293.

On relationship between liver histology and IPF, liver biopsies were done in ten of 15 patients with IPF before diagnosis of IPF. The period before diagnosis of IPF was 10.2 ± 4.1 years. There was no evidence of plasma cells to indicate possible autoimmune hepatitis, which as a systemic disease could be associated with IPF.

In the HCV-group, 15 patients developed IPF during a mean observation period of 8 years. The cumulative rate of newly diagnosed IPF was 0.3% at the end of the 10th year, and 0.9% at the 20th year (Figure 1). On the other hand, none of the patients developed IPF during a mean observation period of 6.3 years in the HBV-group. The cumulative rate of newly diagnosed IPF in the HCV-group was higher than that in the HBV-group (P = 0.021).

We then investigated the factors, except for virus marker, associated with the incidence of IPF in all the 8200 patients with HBV or HCV (Table 3). Univariate analysis identified the following three factors that influenced incidence of IPF: age (P < 0.001), state of liver disease (P = 0.007), and smoking index (P = 0.015).These three parameters were entered into multivariate Cox proportional hazard analysis. The IPF development rate of HCV positive patients was high with statistical significance in the following cases: (1) patients ≥ 55 years (P < 0.001); (2) patients who had smoking index (package per day × year) of ≥ 20 (P = 0.002), (3) patients who had liver cirrhosis (P = 0.042).

Next, we examined the factors associated with the incidence of IPF in all the 6150 patients with HCV multivariate Cox proportional hazard analysis. The IPF development rate of HCV positive patients was high with statistical significance in the following cases: (1) patients ≥ 55 years (OR: 14.24; 95% CI = 3.39-59.74, P < 0.001); (2) patients who had smoking index (package per day × year) of ≥ 20 (OR: 5.43; 95% CI = 1.74-16.88, P = 0.003); (3) patients who had liver cirrhosis (OR: 3.82; 95% CI = 1.23-11.86, P = 0.02).

Table 4 summarizes the characteristics of 15 patients who developed IPF in the HCV-group. Thirteen patients (12 men and 3 women; median age at the time of onset of IPF = 70 years, range = 57 to 79 years) with IPF were diagnosed by well-trained respiratory specialists.

During the observation period, 12 of the 15 patients with IPF in the HCV-group died. Seven patients died of liver-related disease (HCC, decompensated liver cirrhosis, rupture of esophageal varices) and five patients died of IPF. Liver-related death accounted for 58% (7/12) of all deaths and HCC was the major cause of liver-related deaths. IPF-related death accounted for 42% (5/12) of all deaths. On the other hand, of 6002 HCV positive patients without IPF, 1905 died of various diseases during the observation period. Of 1905 patients who died, 44 patients died of lung-related disease, such as acute pneumonitis or pulmonary tuberculosis. The incidence of death based on lung-related disease in patients with IPF was significantly higher than that in patients without IPF (P < 0.001).

Idiopathic pulmonary fibrosis (IPF) is present in patients with chronic hepatitis C virus (HCV) infection. In any case, there is little or no information on the yearly cumulative incidence and risk factors on the development rate of IPF in patients with HCV.

A few previous studies have presented conflicting results with some suggesting that an incidence of anti-HCV antibody positivity in patients with IPF is significantly higher than that in patients without IPF both in Italy and Japan, whereas others found that an incidence of anti-HCV antibody positivity is not high compared to controlled patients.

The morbidity for IPF was estimated to be 0.003% to 0.004% in the general population in Japan. This retrospective study is the first to determine the annual incidence of IPF among patients with HCV at 0.03%-0.04%. The results indicate that age, liver cirrhosis and smoking enhance the development of IPF in patients with chronic hepatitis C infection.

The fact that patients with IPF have a high possibility of dying from acute exacerbation due to IPF during the follow-up shows the need to provide a high level of care to patients with IPF. In general, hepatologists regard the daily management of patients with HCV. When HCV patients complain dry cough and dyspnea, hepatologists should check the complication of IPF.

IPF was diagnosed by respiratory specialist based on the diagnostic criteria recommended by American Thoracic Society/European Respiratory Society. Smoking index was defined as (package/d) × year, indicating the sum of before and after first consultation; period after IPF development, period (year) between onset of IPF and final consultation.

The manuscript is well written and the study is well designed. Authors investigated the cumulative development incidence and predictive factors for idiopathic pulmonary fibrosis in HCV positive patients.