Natural History, Clinical Manifestations, and Prognostic Indicators of Disease Progression and Survival of Hepatitis C Virus (HCV) Infection
July 9-14, 2000
The most accurate means of determining the entire spectrum of outcomes of HCV infection would be to identify a large cohort at a time of initial infection; to select a carefully matched seronegative control group who, like the seropositive group, could be maintained under active surveillance for periods of 30 years or more; and to omit therapy that might alter outcome and hence modify natural history. Obviously, these conditions cannot be met and, therefore, the desired information may never be forthcoming.
Numerous HCV natural history studies have attempted to detail the course of HCV infection, yet most have been unable able to chart disease progression from initial infection to end-stage disease, and only one has more than 25 years of follow-up (Seeff 2000b). Long-term, accurate natural history data are needed to better understand the pathogenesis of the disease and to determine which patients need treatment now, which ones can wait for more effective treatment, and which ones will probably never need treatment.
Clinical Features and Outcome of Acute HCV
Acute HCV infection most often exhibits no clinical symptoms: 60% to 75% of individual are asymptomatic; 20% to 30% become jaundiced; and 10% to 20% have symptoms including fatigue, nausea, and vomiting (Dienstag 1983; Aach 1991; Koretz 1993). In those who develop jaundice, peak bilirubin levels are usually less than 12 mg/dl (mean: 8.4 mg/dl) and levels appear to resolve in less than four weeks (Esteban 1999). Fulminant hepatic failure with primary HCV infection is extremely rare.
Approximately 20% of newly individuals have symptoms that arise before the seroconversion to anti-HCV occurs (antibody-positive) (Koretz 1993). The average time from exposure to seroconversion is approximately 50 days, although it can be as long as nine months (Tremolada 1991; MJ Alter 1992).
During acute infection, a person's aminotransferase (ALT) levels (liver enzymes) rise to 200600 IU/I, but in 20% of cases, peak ALT levels exceed 1000 IU/l (Esteban 1999). There is often an episodic, fluctuating pattern of ALT levels during the first few months, where levels flare to a 10- to 15-fold increase. Approximately 25% of patients develop a sustained plateau pattern where ALT levels remain below 450 IU/l for many months. In others, ALT level can normalize (suggesting recovery), yet flare up again within months to years. This pattern is a tell-tale sign of chronic infection with HCV and indicates a need for ongoing ALT monitoring (MJ Alter 1992; Esteban 1999). More than 12 months after acute infection, ALT levels will continue to be elevated in about 60% of individuals (Esteban 1998).
Continued normalization of ALT levels -- termed "biochemical recovery" -- does not always mean a loss of anti-HCV or absence of HCV RNA (MJ Alter 1992). Likewise, continued abnormal ALT levels with anti-HCV positivity does not necessarily mean that a person is chronically infected with HCV. Fifteen to twenty-five percent of acutely HCV-infected individuals will clear their infections (Shakil 1995). Complete resolution of HCV infection is defined as both the absence of HCV RNA in serum and a normalization of serum ALT level. Shakil and colleagues from the National Institutes of Health (NIH) documented the rate of chronic HCV, using PCR and bDNA assays, and histologic (liver cell) damage in 60 anti-HCV positive individuals with chronic hepatitis (Shakil 1995). Tabled below are the results of the 60 patients equally divided into three groups: group 1: normal ALT levels; group 2: elevated ALT levels less than twice the normal range; and group 3: levels more than twice the normal range.
Some individuals (~15%) who test HCV RNA-positive during acute or post-acute HCV infection may eventually become undetectable and, conversely, some (~19%) who initially test HCV PCR-negative may become detectable at a later date. Villano and colleagues from Johns Hopkins studied HCV RNA patterns in 43 HCV seroconverters (documented by EIA-2 and RIBA-2) who were followed for 72 months (Villano 1999). Six (14%) patients had documented viral clearance (undetectable HCV RNA, <500 copies/mL) between one and two years; one patient was initially undetectable for HCV RNA and remained negative. Two of the patients who cleared virus had ~1 million copies/mL of virus. Factors associated with viral clearance were: white race (P = 0.004); jaundice (P = 0.003); and lower peak viral titers (P = 0.003).
The presence of HCV RNA did not always precede the development of antibodies: 48% had HCV RNA at their first seroconversion visit; 33% were HCV RNA-positive a median 3.8 months before seroconversion; and 19% were HCV RNA-detectable a median 15.3 months after estimated date of seroconversion. The fact that a person who may initially test qualitatively HCV RNA-negative (no detection of HVC in serum) but later become positive suggests that "in clinical practice, virologic outcome must be determined by long-term follow-up, not a single HCV RNA level" (Villano 1999).
One significant finding in this study deserves comment. Seventy-four percent of the anti-HCV-positive patients in this study (the majority with a history of intravenous drug use (IVDU)) had been evaluated by a physician during their seroconversion intervals, yet only a few were recognized as having HCV or were screened for it. In this case, one cannot say that hepatitis was not detected because of limited access to health care. Even though HCV does not present with severe symptoms, this routine failure to diagnosis HCV in patients known to be at high risk indicates a striking lack of awareness by health care providers about HCV and its epidemiology.
Data from various studies indicate that 75-85% of individuals with acute HCV infection will become chronically infected (MJ Alter 1992; Shakil 1995; Villano 1999). The reason for such a high rate of chronicity is not completely understood. Some studies suggest that HCV persistence is related to the high mutation rate of HCV and the continual turnover of complex viral quasispecies that are able to evade the immune response of the host (Tsai 1998; Ray 1999). Farci and colleagues from the University of Cagliari and Robert Purcell's NIH laboratory recently published data which suggest that resolution of acute hepatitis is associated with relative evolutionary stasis of quasispecies, but progression to chronicity is correlated with genetic evolution of HCV (Farci 2000).
Only 60% to 70% of chronically infected individuals will have persistent or fluctuating ALT elevations; the other 30% to 40% will have normal ALT levels throughout the course of their infection. Two studies, one by Shakil and colleagues, the other by Pastore and colleagues have documented that relatively low (13-135) or normal ALT levels during acute infection were prognostic indicators predicting loss of HCV infection (Pastore 1985; Shakil 1995). In at least four other natural history studies, however, ALT levels were not found to be predictive of clearance (Mattsson 1989; Esteban 1991; MJ Alter 1992; Villano 1999). Except for the Villano study mentioned above, which noted that whites had a higher rate of clearance than blacks (80% of the subjects were black), and a small Swedish study (N = 37) by Mattsson that documented an increased rate of clearance among patients under 30 years old, no other major natural history studies have found any prognostic factors for disease clearance (Mattsson 1989; Villano 1999). In 2000, it is still a guessing game as to who will develop chronic infection. According to Juan Ignacio Esteban, "No clinical, biochemical or virological feature can predict the outcome of infection in a given individual" (Esteban 1999).
A liver biopsy gives the most accurate information about the degree of liver injury associated with HCV infection. There are many complex and detailed staging and scoring systems for liver biopsies (e.g., Knodell score, METAVIR fibrosis score).
According to Leonard Seeff, natural history studies have used one of three evaluation strategies (Seeff 1997, 2000a):
One of the most important prospective natural history studies, the NIH Prospective Study of HCV-Infected Donors, was recently published (HJ Alter 1997). It included 280 HCV RNA- positive blood donors followed for a median of 20 years. ALT levels were repeatedly normal in 17%; 45% had <2X the upper limit of normal; and 38% had a least one reading 5X the upper limit. Liver biopsies were performed on 81 of 280 patients, and a probable date of exposure could be ascertained in 74 of the 81. The chart below represents biopsy findings according to time from exposure:
Only 1.3% had histologic evidence of severe hepatitis and cirrhosis following a mean interval of 18 years from time of exposure. According to Harvey Alter, "Liver-related mortality or severe morbidity is less than 10% in the first two decades after infection." (Alter 1997)
Koretz and colleagues from UCLA documented a more progressive clinical course in their study of 80 recipients of HCV transfused blood who were followed for 14 years (Koretz 1993). After 16 years of known infection, 10% had HCV-related symptoms; 18-20% had biopsy-proven cirrhosis; 1.3% developed hepatocellular carcinoma (HCC; liver cancer); and 2.5% died of liver-related complication.
Cohort studies of chronic HCV patients referred to liver-oriented tertiary-care centers suggest progressive HCV disease. Tong and colleagues followed 131 patients with HCV for approximately 22 years: mean age at transfusion was 35 years, and at time of evaluation, the mean age was 57 years (Tong 1995). After about four years of follow-up, over 67% were experiencing HCV-related symptoms (specifically fatigue); 46% had biopsy-proven cirrhosis; 10.6% had developed HCC; and 15.3 % had died of liver-related deaths. The accompanying chart documents time to hepatic event:
Two significant retrospective natural history studies were recently published reporting the longest follow-up yet of well-characterized HCV infected cohorts (Kenny-Walsh 1999; Seeff 2000b). The first, by Kenny-Walsh and colleagues, documents the clinical course of HCV in a group of 376 HCV RNA-positive Irish women who had been infected during 19771978 with a batch of HCV-contaminated anti-D immune globulin (Kenny-Walsh 1999). The mean age of the women at time of infection was ~28 years, and after 17 years of living with HCV, the mean age of at screening/biopsy was 45 years. None had received any anti-HCV treatment. HCV-related symptoms, histologic grade of hepatic inflamation, and stage of fibrosis for the 363 women who underwent biopsies are documented below:
The 45-year HCV natural history study of 8,568 military recruits by Seeff and colleagues reports the earliest confirmed detection of HCV in the United States and has the longest follow-up of any study published to date (Seeff 2000b). Originally, 8,568 military recruits were tested for group A streptococcal infection and acute rheumatic fever between 1948 and 1954. After initial blood tests, samples were frozen and saved for 45 years. Seventeen (0.2%) individuals tested positive for HCV antibodies on EIA-3 and RIBA-3. While 17 HCV-positive patients followed in a natural history study may seem small, it's the HCV-negative control group of over 8,000 individuals (99% men) that makes this study important. Records tracking 45 years of liver abnormalities, disease, and death (using the National Death Index Plus service) as well as age, sex, race, chart review, and cause-specific mortality from death certificates were available from the Veterans Affairs' computer files.
A vast majority of the recruits were younger than 25 years of age at the original blood draw. Approximately 90%, 10%, and 0.5% were white, black and Asian, respectively. HCV infection in blacks was significantly higher than whites: 1.8% compared to 0.07% (RR, 25.9; CI, 8.4 to 80.0). The mean age of the surviving cohort, as of January 1997, was 64.8 years (95% were between 60 and 69 years of age). Eleven of the seventeen (65%) HCV-infected men were HCV RNA-positive; all but one had genotype 1b.
Prognostic Factors for Fibrosis Progression:
The mean and median rate of fibrosis progression per year for the 1,157 patients whose duration of infection was known was 0.252 (95% CI, 0.227-0.277) and 0.131 (95% CI, 0.125-0.143), respectively. At this rate the median time to cirrhosis was estimated as 30 years (range: 28-32 years); 33% had a median time of 20 years and 31% will never progress or will not progress for at least 50 years. Out of the nine factors assessed, three had highly significantly correlation with disease progression: gender, alcohol consumption, and age at infection. Interestingly, amount of HCV RNA and viral genotype (i.e., 1b) did not correlate significantly with disease progression. (See "HCV Treatments" chapter, where both are significant prognostic factors for success of treatment.)
Selected Prognostic Factors for Disease Progression and Survival
Age (Tremolada 1992; Tong 1995; Fatovitch 1997; Poynard 1997; Neiderau 1998; Deuffic 1999); male gender (Poynard 1997; Deuffic 1999; Khan 2000); and increased alcohol intake (Donato 1997; Fattovitch 1997; Poynard 1997; Rudot-Thoraval 1997; Neiderau 1998; Wiley 1998) have been documented by many (but not all) as highly prognostic factors for HCV-related liver disease progression. However, other important prognostic factors for advanced disease progression have been less consistently documented in published natural history studies. Mode of HCV transmission: Blood transfusion vs. IVDU vs. Sporadic Cases.
Two published studies have suggested that infection through blood transfusion (BT) (compared to infection via IVDU) is a highly significant prognostic factor for liver disease progression (Rudot-Thoraval 1997; Gordon 1998). Gordon and colleagues studied the clinical course of 627 chronically HCV-infected patients: 282 (45%) were BT recipients, 262 (42%) were infected via IVDU, and 83 (13%) were without known risk factors. The median estimated disease duration for all patients was 21 years (+/- 9.53 years) and the duration of follow-up ranged from 1 to 25 years. Liver histology was available on 463 patients. Cirrhosis was determined in 173/463 (37%): 118/173 (68%) were BT recipients; 40/173 (23%) were infected via IVDU (P<0.001). Below is a breakdown of those who had HCC, cirrhosis, or no cirrhosis:
Unlike the Poynard study (discussed above), in Gordon's patients, age or estimated disease duration did not predict risk of liver failure in the multivariate analysis (Poynard 1997; Gordon 1998). Rudot-Thoraval and colleagues from France also noted an increased prevalence of cirrhosis in BT patients in their survey of 6,664 chronic HCV patients (Rudot-Thoraval 1997).
Of 2,500 patients with known duration of HCV infection, the prevalence of cirrhosis for BT recipients was 22.8% compared to 5.8% for those infected via IVDU (P<0.02; OR = 0.61; 95% CI, 0.40-0.92).
"Sporadic cases" are HCV infections without identified risk factors. Fattovitch and colleagues from Italy noted that sporadic cases with compensated cirrhosis had poorer survival compared with those infected via BT or IVDU (Fattovitch 1997). Likewise, Khan and colleagues from Australia recently reported that sporadic cases were at significantly higher risk for cirrhosis, HCC, and liver transplantation or death (Khan 2000). The complete opposite was seen in a German HCV natural history study conducted by Hopf and colleagues (Hopf 1990). Finally, Poynard did not find any of these modes of transmission to be a significant prognostic factor in his study (Poynard 1997).
Abnormalities in Laboratory Values: Albumin and Bilirubin
End-stage disease from hepatitis C is one of the leading indications for liver transplantation in the United States. Currently, listing for transplantation requires significant abnormalities in at least 2 of 5 elements of the Child-Pugh [cirrhosis] classification. This may merit reassessment if the dramatic predictive value of the serum albumin alone can be validated in other large, prospective studies. (Hirsch 2000)
To a lesser extent, elevated bilirubin has been noted as a risk factor for HCV-related liver disease progression (Fattovitch 1997; Khan 2000). In their 1997 study of 384 Italians with cirrhosis, Fattovich and colleagues found that abnormally high bilirubin was a predictor of poorer survival (Fattovitch 1997). Those with bilirubin <17 mmo/L had five- and ten-year survival probabilities of 96% and 86% respectively, compared with 81% and 67% for those with bilirubin >17-<51 mmo/L (P = 0.0001).
Prolonged prothrombin time (decreased duration of blood coagulation) and decreased platelet count have been documented as significant predictors of disease progression in later-stage patients (Fatovitch 1997; Gordon 1998).
Coinfection with HAV, HBV, and HIV
(See "Hepatitis & HIV Coinfection" chapter for more details)
HCV patients who acquire HAV have been found to have a substantial risk for fulminant hepatic failure2 and death (Vento 1998). In an Italian HCV natural history study, 432 patients were tested thrice yearly for the development of HAV antibodies. Seventeen HCV patients (three with cirrhosis) subsequently acquired HAV infection. Ten of the patients had an uncomplicated course of HAV, but seven developed fulminant hepatic failure, and six died. There was no apparent difference in degree of baseline liver damage, yet the development of HAV posed a 41% chance of fulminant hepatitis and a 35% chance of death (Vento 1998).
Vento concludes that all HCV-infected individuals should be vaccinated for HAV, saying, "Chronic carriers of HCV who are at risk for HAV infection should be vaccinated against HAV, since superinfection with this virus may place them at risk for severe, life-threatening acute liver damage" (Vento 1998). In a subsequent Lancet editorial, Marina Berenguer and Teresa Wright agree that HAV vaccination may be warranted, but believe that Vento's results need confirmation and that a cost-benefit analysis is essential before implementing such a policy worldwide (Berenguer 1998).
HCV patients with HBV have been found to have an increased risk for cirrhosis (Roudot-Thoraval 1997; Cacciola 1999). Cacciola and colleagues from Italy tested for the presence of HBV DNA in a cohort of 200 chronically HCV-infected individuals. Sixty-six (33%) were found to have HBV sequences. Of these 66 patients, 22 (33%) had evidence of cirrhosis compared to 26 (19%) of 134 HCV-positive/HBV-negative patients (P = 0.04). This study confirmed the results of a French study of 5,786 histologically HCV-diagnosed patients (Roudot-Thoraval 1997). Cirrhosis at liver biopsy was found in 24.6% of patients positive for HBV surface antigens compared to 21.1% in those who were HBV-antigen-negative (P<0.001; OR = 1.99; 95% CI, 1.40-2.82).
Studies of people coinfected with HCV and HIV have reported that, while the progression of HIV disease is not changed by HCV, HCV infection progresses more rapidly in people with HIV. When matched for other variables, on average, people who are HIV-positive have higher levels of HCV RNA than HIV-negative people (Eyster 1993, 1994; Cribier 1995). It is important to note that a majority of natural history coinfection studies were conducted before the advent and widespread use of potent antiretroviral therapy. With control of HIV viremia and better immune status, the natural history of HCV in these patients will likely change. In immunocompetent HIV-positive individuals, HCV may replace certain opportunistic infections (i.e., Pneumocystis carinii pneumonia, mycobacterium avium complex and cytomegalovirus retinitis) as a leading cause of increased morbidity and mortality.
Benhamou and colleagues from Theiry Poynard's group in France have recently reported on fibrosis progression in their well-characterized coinfected DOSVIRC cohort (Benhamou 1999a, 1999b). Low CD4 count (<200 cells/mm3), alcohol consumption of more than 50 grams/day, and age at HCV infection (>25 years old) were shown to be associated with an increased liver fibrosis progression rate (Benhamou 1999b). In a linear progression model, there was little difference in time from infection to cirrhosis between the HCV+/HIV+ patients and the matched HCV+/HIV- groups if those with coinfection who had a T-cell count of >200 and drank <50 grams/day of alcohol. The accompanying chart documents time to cirrhosis:
Quantitative HCV RNA levels do not appear to affect the clinical course of HCV-infected individuals who are naive to treatment. Lau and colleagues found no difference in serum HCV RNA levels between patients with chronic persistent hepatitis, chronic active hepatitis, or cirrhosis (Lau 1993). These finding were confirmed in larger natural history studies (Poynard 1997; DeMoliner 1998).
There is considerable debate as to whether certain HCV genotypes, especially 1b, put patients at increased risk for disease progression. Kabayashi and colleagues found that the deterioration of liver histology during a median 9.6 years of follow-up was more common in patients with genotype 1 (68%) -- namely 1b -- than in those with genotype 2 (41.7%) (P<0.01) (Kobayashi 1996). Likewise, an advanced stage of liver histology was more common in the genotype 1 patients (63%) compared to those with genotype 2 (38.9%) (P<0.05). An earlier, smaller study conducted in the United Kingdom also noted that those with genotype 1 had a far more progressive histologic disease than those with genotypes 2, 3, or 4 (Dusheiko 1994). However, difference in disease outcome according to genotype has not been verified in most treatment-naive natural history studies after 1996 (Poynard 1997; Serfaty 1997; Neiderau 1998; Khan 2000). Finally, in interviews conducted with leading hepatology researchers and clinicians, none believe that genotype independently has an effect on the natural history of hepatitis in patients naive to therapy (see "Current Opinions and Controversies" chapter).
Aach RD, Stevens CE, Hollinger FB, et al. Hepatitis C virus infection in post-transfusion hepatitis: an analysis with first- and second-generation assays. N Engl J Med 325:1325-9, 1991.
This article was provided by Treatment Action Group. It is a part of the publication The Hepatitis Report.