Hepatocellular carcinoma in human immunodeficiency virus (HIV)-infected patients: Is it really different, and if so, why?☆

1.1. Predictive factors for HCC development 

When discussing predictive factors for HCC in HIV-infected patients, the role of highly active antiretroviral therapy (HAART) as a moderator of HCC risk should be considered. One study [12] showed that the incidence of cirrhosis in HIV and HCV co-infected patients increased from the pre-HAART to the HAART era, a result that may imply that the improved survival due to HAART led to longer exposure to HCV and hence a higher risk of cirrhosis; however, it should be taken into account that the risk of cirrhosis was still higher after adjusting for survival time. Other factors in association with hepatitis virus could increase the risk of HCC. As in monoinfected patients, heavy alcohol intake will favor the development of cirrhosis and HCC and the severity of liver disease can also be boosted by any drug-induced hepatitis. Some studies have found an association between obesity and the risk of HCC in the general population [13]. Insulin resistance is strongly associated with obesity and contributes to liver steatohepatitis, which is associated with more severe liver necroinflammatory activity, fibrosis and cirrhosis [14]. The link between insulin resistance and the development of HCC has been established [13], and it might be a risk co-factor in the HIV setting due to high rate of insulin resistance induced by HAART [15].

Although there is currently some confusion surrounding the role of HAART in accelerating or preventing the development of HCC in HIV, the beneficial effect of HAART in slowing liver fibrosis progression and improving survival in HIV and HCV co-infected patients should be kept in mind [16]. Moreover, a recent study that aimed at estimating the future disease burden of HCV and HIV infection in the United States concluded that the predicted increase HCV mortality can only be tackled by better access to antiretroviral therapy or antiretroviral therapy with greater efficacy [2].

1.2. Clinical characteristics 

HIV infection is known to influence the course of both HBV and HCV infections. In HBV infection it impairs the quantity and quality of the innate and adaptive immune response, leading to a more rapid progression of liver fibrosis and a higher incidence of decompensated cirrhosis (although incidence of HCC is not clearly affected) [17]. Early initiation of HAART could help prevent severe immune dysfunction.

HIV significantly modifies the course of HCV infection, increasing the risk of cirrhosis and accelerating fibrosis progression leading to cirrhosis at an earlier age (12–16years) [18]. This phenomenon leads to the more rapid development of cirrhosis, found in the non-tumoral liver in most cases of HCC. Moreover, the up-regulation of the Fas ligand by HCV E2 and HIV glycoprotein gp120 in HepG2 cells could enhance the formation of the death-inducing signaling complex in fas-induced apoptosis downstream of Fas receptor activation, thereby favoring carcinogenesis [19]. In the two large series of patients with HCC in HIV and HCV co-infected patients published to date [1], [4], the characteristics of HIV co-infected patients were similar, that is, they were younger than non-HIV patients (by 10years), aetiology was similar – three-quarters had HCV, and one-quarter had HBV associated tumour – heavy alcohol consumption was documented (32–46%), the personal history of most patients included drug abuse (75–85%), and HCC occurred in patients with relatively preserved liver function (Child A and B). The characteristics of the tumours were also similar: 35–44% were solitary tumours, with a median size of 4.2cm, and 32% patients had portal thrombosis. The median level of serum alfa-fetoprotein (AFP), considered by some authors as indicative of poor prognosis, was found higher than in monoinfected patients in both series. Overall survival was similar in both series – approximately median survival of 6 months – which is considered a very poor prognosis. This prognosis might reflect an advanced stage of the disease in patients not receiving surveillance but with a natural history that parallels HCC in non-HIV individuals. Alternatively, the poor outcome might reflect a more aggressive biological pattern. Prospective studies are required to elucidate this issue.

In these series, 50–60% of the patients received unproven or ineffective therapy, whereas more aggressive treatments, based on tumour and liver disease characteristics, are worth trying. None of the patients was considered for liver transplantation, which is no longer contraindicated in HIV patients. Although the tumour stage was similar in HIV and non-HIV patients, only a few HIV patients received potentially curative therapies [20], [21], [22] compared to non-HIV patients in both series.

Clearly, clinicians should be more willing to consider more aggressive surveillance and application of effective therapies in this subset of patients, who have a very poor survival at present. Although the HIV-infected patients included in the two series show similar clinical patterns, very different conclusions are drawn from the two studies. In the US-Canadian series [1], patients with HCC had a poor prognosis (median survival of 6 months), and optimal screening was not defined. Indeed, survival curves were similar for HIV-positive and HIV-negative patients or HCV monoinfected and HIV/HCV co-infected patients. The European study [4] found a median survival of 700–725days in monoinfected patients in both series (Brescia HCC study and CLIP database). Results are however difficult to compare because of the lack of a common staging system used. In the US–Canada cohort more than half of the patients belonged to BCLC C or D. In this stage of the disease, the natural history of untreated individuals is of a median survival of 6 and 3 months, respectively [23], [24]. Conversely, in patients at intermediate stage of the disease (BCLC-B) the median survival of untreated patients is 15 months [23], [24]. Thus, differences in outcome might reflect differences in staging between both studies. The reasons for these discrepancies in HCC outcome require closer investigation. Is the difference the result of a more aggressive detection policy in Europe? Is it because the outcomes of effective therapies, namely surgery, radiofrequency and chemoembolization, are different? The former is more likely.

In conclusion, HCC in HIV patients (generally co-infected with HCV) is currently detected at advanced stages of the disease, and thus prospective detection according to reported surveillance programs is recommended [25]. Second, we should consider all types of HCC treatment not just for HIV-negative patients but also for HIV co-infected patients, given that the very poor prognosis in HIV co-infected patients may be improved by both aggressive therapeutic options and best supportive care. In addition, with the advent of sorafenib as an efficacious treatment of HCC in advanced stages [26], prospective studies assessing the effects of this treatment in HIV patients are warranted.