Increased Incidence of Cancer Observed in HIV/hepatitis C Virus-Coinfected Patients Versus HIV-Monoinfected

Héctor Meijide; Sonia Pértega; Iria Rodríguez-Osorio; Ángeles Castro-Iglesias; Josefa Baliñas; Guillermo Rodríguez-Martínez; Álvaro Mena; Eva Poveda


AIDS. 2017;31(8):1099-1107. 

In This Article


In 2318 PLWH followed for 26 580 person-years, 8.0% of patients developed at least one cancer. This is in accordance with other previously published data, such as a large French study of 99,817 PLWH followed for 18 years in which 7.7% presented with at least one tumor,[13] or the recently published data from the Veterans Aging Cohort Study (VACS) from North America, in which the 7.8% of the PLWH developed at least one cancer.[28] In our cohort, PLWH has double risk of developing a NADC than the general population, after adjustment for sex and age.

The analysis revealed a higher incidence of ADC in HIV-monoinfected than in coinfected patients. KS accounted for approximately 50% of ADC in HIV-monoinfected patients (41.3%), but only 12.5% of the ADC in coinfected patients. KS affects mainly to MSM, and the proportion of MSM in the coinfected group was lower in our cohort. KS is frequently an early manifestation of HIV infection, sometimes the first one; in this study, some coinfected patients were diagnosed with HCV infection months or years after HIV infection, and they were considered monoinfected up to the HCV-coinfection diagnosis. Active drugs users had less access to the health system during the 1990s; this could also contribute to the lower incidence of ADC in coinfected versus monoinfected patients during the first period (1993–2003).

The timing of cancer incidence after HIV diagnosis differed between ADC and NADC, being much shorter in ADC as shown previously.[12] However, in this study, we found that HIV/HCV-coinfected patients developed both ADC and NADC significantly later than HIV-monoinfected patients. Coinfected patients were mostly IDUs, with a higher mortality due to other reasons (violence, substance abuse, etc.), which could compete with the risk of early cancer development.

The competing risk analysis showed that, after adjusting for age at HIV diagnosis, sex transmission route and considering death without cancer a competitive risk; coinfected patients had a higher cumulative incidence of NADC than HIV-monoinfected. The weight of HCC in the incidence of NADC in the coinfected group is unquestionable but, when the HCC is excluded from the analysis, the cumulative incidence of NADC remains significantly higher in coinfected than in monoinfected patients (adjusted SHR = 1.26). HCV infection appears to have a role in cancer incidence and mortality and a higher incidence of lung, digestive tract and kidney cancers in HCV-infected patients has been reported previously. The molecular mechanism is unknown, but extrahepatic manifestation of hepatitis C with a chronic inflammatory condition, such as glomerulonephritis, cryoglobulinemia or lichen planus, can play a role.[29–31] HCV eradication reduces liver-related and nonliver-related mortality in patients with chronic HCV infection.[32] Nevertheless, the specific role of HCV eradication in cancer development is unknown.

The HCC represents the 34.6% of the NADC in HIV/HCV-coinfected patients. In our study, the HCC incidence in coinfected patients was 1.42/1000 person-years, close to that reported in other recent study with a mainly European population (1.59/1000 person-years)[33] but below the 4.44/1000 person-years published in a huge US Veterans cohort.[17] The older age, ethnic variability and earlier HIV-infection acquisition in the US cohort may explain this difference. Gjærde et al.[34] reported an alarming increase in HCC incidence between 2001 and 2014 (up to 2.3/1000 person-years); in our cohort, all HCCs were diagnosed after 2003 and the incidence will presumably continue to grow over the following years. The development of new drugs against HCV and the eradication of HCV co-infection in PLWH may improve the prognosis of coinfected patients. Nonetheless, the impact on HCC incidence is not well established. Recent data suggest the increased risk of HCC early recurrence in patients treated with direct-acting antivirals for HCV infection, but more studies are needed to confirm these preliminary data.[35,36]

A recent study by the Collaboration of Observational HIV Epidemiological Research Europe (COHERE) has demonstrated a reduction of NHL incidence in monoinfected PLWH, but this favorable impact is minimized in hepatitis-coinfected patients.[24] In our study, the incidence of ADC in HIV-monoinfected patients decreases with the time, probably related with ART exposure and immune restoration, but the incidence in coinfected (most of cases NHL) is increasing, despite the ART exposure, in accordance with the COHERE Study, where HIV/HCV-coinfected patients receiving ART had higher risk of NHL than HIV-monoinfected patients (HR = 1.73). Several studies in HIV-uninfected persons have investigated the association between chronic HCV and lymphoma development, with odds ratios between 2.0 and 2.5.[37] All these data suggest that HCV-coinfection has a role in the development of NHL, which becomes more evident when controlling the immunosuppression caused by the HIV-infection. Diverse mechanisms have been proposed, such as a direct oncogenic role of HCV, an immune dysregulation or a more prevalent coinfection by other viruses involved in lymphoma development (mainly herpes viruses). Without the eradication of viral hepatitis, it is expected that the incidence of NHL in coinfected will increase.

The immune status is closely related to the development of ADC. With the introduction of ART, many studies have observed a progressive decline in the incidence of ADC.[28,38,39] The inverse relation between CD4+ cell count and NADC incidence is controversial. Some studies have documented the impact of a higher CD4 and ART exposure on lower NADC incidence (mainly in virus related-NADC), whereas others have not.[28,40,41] We found differences in CD4+ cell count and ART exposure between HIV-monoinfected and HIV/HCV-coinfected patients at cancer diagnosis, but these differences were related to the higher incidence of ADC in the monoinfected group. No differences between mono- and coinfected patients were found when patients with NADC were compared. A recent analysis from the START study, published by Borges et al.,[42] demonstrated that immediate ART initiation in patients with more than 500 CD4+ cells/μl significantly reduces risk of cancer, mainly at the expense of infection-related cancer.

This study has several limitations. First, the retrospective design cannot ensure that all patients with cancer were included, mainly those with ambulatory treatment (skin tumors or high-grade cervix lesions). However, the regular monitoring in the HIV unit and the low mobility of the patients minimized this bias. Data regarding tobacco and alcohol consumption were not available, and we were unable to control for them in the multivariate analysis, and presumably their prevalence was higher in the coinfected group, which would be an important confounder. The cancers in this study were classified as ADC and NADC; currently, some studies use the infection-related and unrelated cancers classification for grouping cancers as infections and immunosuppression could play a role, but this classification also has some inaccuracies. In addition, the treatment data and anatomic stage from the registries, particularly in the first years of the study period, may have been incomplete. GLOBOCAN provides estimates of cancer incidence, mortality, and prevalence for countries and world regions. Incidence data are derived from population-based cancer registries that vary in coverage and may capture the population of an entire country, but more often cover smaller areas, such as mayor cities, plus it is a point in time record and, in this study, it is used comparatively with a cohort followed for 22 years; despite these limitations, GLOBOCAN 2012 is a key source of information on the profile of cancer and represents the best estimates available. Finally, comparison with GLOBOCAN in persons older than 64 years must be interpreted with caution as this age group is under-represented in our cohort (3.45% globally). Nevertheless, it is expected than the incidence increases with the progressive aging of PLWH.

In conclusion, malignancies are an important comorbidity in PLWH, with a higher incidence than in the general population. After adjusting for epidemiological factors and mortality without cancer, HIV/HCV-coinfected patients presented more NADC than HIV-monoinfected patients, even excluding the HCC. Treatment of HCV infection and HIV replication control are fundamental strategies, but the valuable role of cancer-screening programs and early treatment must be assessed.