HIV Articles  
Emerging role of hepatocellular carcinoma among liver-related causes of deaths in HIV-infected patients: The French national Mortalite 2005 study
  Jnl of Hepatology April 2009
Dominique Salmon-Ceron12, Eric Rosenthal34, Charlotte Lewden56, Vincent Bouteloup5, Thierry May78, Christine Burty7, Fabrice Bonnet69, Dominique Costagliola1011, Eric Jougla12, Caroline Semaille13, Philippe Morlat569, Patrice Cacoub1114, Genevieve Chene569, the ANRS EN19 Mortalite Study Group and Mortavic
Longer exposure to hepatitis C (HCV) or B virus (HBV) and the increased use of hepatitis treatment might have an impact on liver-related deaths in patients co-infected with the Human Immunodeficiency Virus (HIV). We describe the proportion of liver-related deaths among HIV-infected patients in 2005 compared with 2000.
In a nationwide survey (341 hospital departments involved in HIV management), all deaths of HIV-infected patients were prospectively reported. Deaths from either cirrhosis, hepatocellular carcinoma or fulminant hepatitis were defined as liver-related deaths.
Of the 898 deaths reported in 2005, liver-related causes accounted for 15.4%; this is compared to 13.4% in 2000. Among liver-related deaths, hepatocellular carcinoma increased from 15% to 25% (p=0.04). Among hepatocellular carcinoma-related deaths: in 2000, 10% were HCV-infected; in 2005, 25% were HCV-infected (p=0.03). Half of the HCV-related deaths had been treated for HCV but 98% remained HCV-RNA positive at time of death. The proportion of HBV-related deaths remained stable between 2000 and 2005.
Liver-related deaths, mainly liver cancers, have increased in HIV-infected patients in France despite wide access to HCV treatment. The stability of HBV-related deaths might be explained by the use of dually active antiretroviral drugs in co-infected patients.
In recent years, end-stage liver disease has become a major cause of morbidity and mortality in patients chronically infected with the Human Immunodeficiency Virus (HIV) who are co-infected with hepatitis C (HCV) or B virus (HBV) [1], [2], [3], [4]. This trend may be explained by prolonged exposure to viral hepatitis, because most patients were infected with HCV in the 1980s. This time interval is compatible with the occurrence of cirrhosis and its late complications, as accelerated fibrosis and cirrhosis are observed in more HCV-HIV co-infected patients compared to HCV mono-infected patients [5], [6].
In parallel, thanks to the widespread use of combination antiretroviral therapy (cART) in industrialised countries, the incidence of Acquired Immune Deficiency Syndrome (AIDS) morbidity and related mortality has dramatically decreased, resulting in an increased life expectancy. However, this has also meant that the causes of death have shifted to predominantly non AIDS-related diseases, with a high proportion of end-stage liver diseases being observed [2], [7], [8], [9], [10]. These major hepatic complications have emerged in the context of excessive alcohol use and metabolic issues such as steatosis [11] and portal hypertension [12]. These complications may also be related to antiretroviral treatment toxicity but their role in liver mortality remains to be defined.
On the other hand, treatments for viral hepatitis are now widely available in industrialised countries and viral eradication is achieved in 25-40% of HCV-HIV co-infected patients treated with a peginterferon and ribavirin combination [13], [14], [15], [16], while anti-HBV treatment is a part of cART. However, the best treatment strategy for HCV in patients with an indication for cART remains controversial.
Several studies including the French national survey, "Mortalite 2000", have shown that complications of HCV and HBV are the second most frequent underlying cause of death after AIDS in HIV-infected patients, accounting for around 10% of deaths [1], [2], [3], [4].
One of the aims of the "Mortalite 2005" survey, was to identify and describe the characteristics of HIV-infected patients who died from end-stage liver disease. These patients were then compared to the population identified in 2000.
In this national survey of deaths occurring in HIV-infected patients in 2005, liver diseases represented the third most frequent underlying cause of death (15.4% versus 13.4% in 2000). This finding has already been underlined in a few surveys such as the D:A:D international collaboration, which reported a similar proportion of liver-related diseases among the causes of death in HIV-infected patients in the cART era, with an estimation of liver-related mortality of 0.24% person-years of follow-up. This echoes the findings of our study which did, however only record deaths [2], [4].
Over the 1995-2003 period, the French Mortavic group performed consecutive surveys that showed a dramatic decline in the proportion of AIDS-related deaths from pre-cART periods (92-46.9%) and an increase in the proportion of end-stage liver disease, as cause of death (1.5-12.6%) [2].
We believe that these results may not be explained by the fact that the wards targeted in 2005 slightly differed from those targeted in 2000, since the proportion of liver-related deaths and patients' characteristics did not differ between wards that participated in both the 2000 and 2005 surveys and wards that participated only in 2005 [17]. Moreover, hepatology wards that participated only in 2005 reported only 3% of all liver-related causes of deaths.
In our study, liver-related death occurred mainly in HCV-infected patients representing 71% of patients with liver-related mortality. In this population, the proportion of hepatocellular carcinoma increased by a factor of 3 over this 5-year period. This increase in the proportion of hepatocellular carcinoma occurred despite better control of HIV infection (median CD4 count at 231/mm3 in 2005 versus 157/mm3 in 2000) and may be the result of a longer exposure to HCV infection. Indeed, although the majority of infections occurred in the late 1980s, the efficacy of cART could have increased survival in cirrhotic patients, which has thus allowed the time for more cases of hepatocellular carcinoma to develop. As all hepatocellular carcinoma cases were symptomatic at the time of death as also shown in a recent study [19], we believe that the recent improvement in the diagnostic accuracy of hepatocellular carcinoma did not play an important role in the increased number of cases observed between 2000 and 2005. The known duration from HIV diagnosis in patients who died from hepatocellular carcinoma, that can be extended to duration since HCV diagnosis, was significantly longer than in patients who died from non-liver-related causes (median: 15.8 years versus 10.8 years). Apart from HCV infection, known risk factors for hepatocellular carcinoma include male gender, age older than 60 years, HBV infection, excessive alcohol consumption, tobacco, diabetes, and high body mass index [19], [20], [21], [22]. Some of these co-morbidities were frequent in our series of HIV-infected patients dying with hepatocellular carcinoma, like HCV infection (69%), chronic HBV co-infection (17%), excessive alcohol consumption (41%), and tobacco use (60%). The important role of hepatocellular carcinoma in HCV-infected patients may also be related to a high rate of treatment failure, especially since genotypes 1 and 4 represented 67% of HCV liver deaths in our study, and efficacy of pegylated interferon and ribavirin combination is known not to exceed 15% in patients with these genotypes [13], [14], [15], [16].
Conversely, the proportion and the number of HBV-related deaths remained low and stable between 2000 and 2005 and the rate of hepatocellular carcinoma as a cause of death among HBV-infected patients did not increase (7/13 cases in 2000 versus 6/16 cases in 2005). Moreover, there was probably an underestimated proportion of delta co-infection in HBV-infected patients (two cases documented among six hepatocellular carcinoma related to HBV infection in 2005). As it is now well known in HBV mono-infected patients, HBV-DNA levels are directly correlated to the incidence of long-term complications of cirrhosis, hepatocellular carcinoma and deaths. In HBV-HIV co-infected patients, anti-HBV active cART makes it possible to achieve suppression of HBV replication in a substantial proportion of patients particularly those who have a complete HIV suppression and a satisfactory immune recovery [23]. Lamivudine has been widely available since 1996, and if there was a beneficial impact of this therapy then this would have been observed before 2000. The design of our study, therefore, did not allow for exploration of the effect of lamivudine. The widespread use of tenofovir, one of the most promising drugs against HBV in these patients, may likely explain this favourable evolution. Moreover, in cirrhotic patients with long-term suppression of HBV replication under tenofovir treatment, improvement in hepatic functions has been observed, creating hope for a potential reversal of end-stage liver disease, and for a survival benefit in this population [24], [25].
Our study has also drawn attention to the emergence of new, albeit rare, causes of liver disease leading to death. Among them, cART was directly or indirectly associated in three cases including one case of didanosine-associated lactic acidosis in a patient with pre-existing cirrhosis, and two cases of portal hypertension. Although cART has significantly improved the prognosis of HIV infection, long-term complications of these drugs are increasingly recognized as significant causes of morbidity and mortality. Non-alcohol fatty liver disease, which can evolve into non-alcohol steatohepatitis, cirrhosis and ultimately hepatic failure, has been recently described as a complication due to multiple factors that co-exist in HIV-infected patients including metabolic abnormalities, chronic inflammation, concurrent infection with HCV [12]. Among them, stavudine and didanosine have proven to be commonly implicated in the occurrence of mitochondrial abnormalities and of lactic acidosis.
More recently, some cases of non-cirrhotic portal hypertension have been described in HIV-infected patients [26], [27], [28], [29]. In such cases, liver biopsy usually showed vascular lesions presenting as nodular regenerative hyperplasia. Patients underwent the significant complications of portal hypertension, such as variceal bleeding and refractory ascites. It has been hypothesised that these lesions were related to either exposure to didanosine [30] and/or intrahepatic microthrombosis, as two studies have shown coagulation test abnormalities like protein S deficiency, elevated homocysteinemia and a constitutional elevation of plasma factor VIII coagulant activity [12]. The protease inhibitor treatment may have played a role in increasing the thromboembolic risk.
In conclusion, while the proportion of deaths related to HBV end-stage liver disease seems stabilized with effective drugs as part of a cART, there is still an increase in deaths related to HCV end-stage liver disease for which therapeutic options remain limited. The incidence of hepatocellular carcinoma is approximately 2-5% per year in HCV-HIV infected patients with cirrhosis. In the absence of new therapeutic options, an exponential burden of deaths from decompensated cirrhosis or hepatocellular carcinoma in HCV-HIV infected subjects is expected in the near future [31], [32]. To avoid such a scenario, anti-HCV treatment should be widely proposed even in patients with cirrhosis provided that they have no decompensation as early HCV viral kinetics (3-months) predicts sustained virological response and allows stopping therapy in cases of insufficient viral load decrease. As in HBV infection, regression of liver fibrosis has been observed in HCV-HIV co-infected patients after treatment with pegylated interferon plus ribavirin [33]. Active promotion of alcohol and smoking cessation programs should be widely implemented. Regular screening for hepatocellular carcinoma should be performed in order to detect small unifocal lesions accessible to surgery. Moreover, evaluation of new therapeutic strategies including new drugs (protease inhibitors, polymerase inhibitors) and high doses of ribavirin in patients who failed to respond to a first treatment as recently demonstrated in the PRESCO trial [34], are urgently needed, and trials must be developed early in HIV-infected patients in parallel to those initiated in HCV mono-infected patients. Lastly, one must remain vigilant about potential new liver complications which may be related to antiretroviral therapy.
  iconpaperstack view older Articles   Back to Top