This nationwide collaborative study describes epidemiological, virological, and clinical patterns related to HBV infection in HIV-positive patients with HBV viremia in the era of fully active antiviral treatments in France. Recommendations from international scientific societies have been published and point to an initiation of treatment at a lower HBV-VL than previously recommended
[15–18]. Thus nowadays, only very few patients should not be considered for HBV-treatment
. Contrary to previous studies, this collaborative work was initiated from a virology laboratory network and patients were randomly selected on the basis of their analytical values.
On a pure technical point of view, our study clearly demonstrates that diffusion of a consensual protocol for HBV-sequencing to multiple laboratories along with a reference sequence library was efficient to generate high quality information regarding HBV genetic variability. We also validated the web-based algorithm proposed on "
http://www.hiv-grade.de" that can be largely recommended for any laboratory dealing with HBV sequence.
Unequal distribution of genotypes according to the mode of HBV acquisition has been previously described in a small study from Perez-Olmeda et al. and more recently in the large EuroSIDA study group with comparable conclusions to our study and to recent work in France
[5, 6, 20]. From the genotype distribution and the epidemiological data, identification of very distinct populations may lead to different clinical managements. For instance, genotype D or A infected individuals are most likely IVDU or MSM, respectively, who have certainly acquired both infections either concomitantly or in a short-time interval. By contrast, most HIV-positive women with genotype E HBV infection have likely acquired HBV perinatally and in highly endemic countries. These different sequential events may have a profound impact on the natural history of the disease, particularly since they occur in patients of different gender, age, and immunological status. Very few data are available on the chance to spontaneously clear the virus in these different situations but one may expect very different outcomes in each of them as recently shown in a large US HIV-infected patient cohort
. With regards to these observations, it is interesting to note relatively mild liver lesions in genotype E infected patients compared to patients infected with genotype A or D. Thus, natural history of chronic hepatitis B during co-infection should be analyzed with scrutiny as it may largely differ according to the population that is considered and may influence treatment options.
HBV-VL is one of the main criteria to initiate antiviral treatment. In our cohort, HBV-VL was not influenced by HBV genotype but by the presence of HBeAg. Contrary to what observed for HBV mono-infected patients, HBeAg is often present (66%) in HIV co-infected patients, as previously reported
[22, 23]. Interestingly, HBeAg-negative patients were less often treated, reflecting the current recommendations not to treat patients with low viremia
Among the 15 genotype G infected patients, all but two were HBeAg-positive, which is surprising as genotype G naturally carries the G1896A stop-codon mutation that prevents HBeAg translation from the pre-C start codon
. Co-infection with another wild-type viral strain whose genome encodes for HBeAg likely explains a rescue for production of this protein, as reported earlier
. Contrary to a recent report from Lacombe et al., we only found a tendency for a more severe liver fibrosis in genotype G infected patients, without statistically significant association
. Reasons for this divergence might be due to size differences in the studied populations and also to missing information regarding liver histology in few cases from our study.
One major finding of the present study is the relative low rate of 3TC-resistance mutations. This observation is somewhat surprising as 3TC, or its derivative FTC, is very often a key component in the HAART backbone. As shown on Figure
1 and as previously demonstrated, 3TC treatment induces very rapid emergence of resistance in HIV co-infected patients
. The distribution of 3TC-resistance mutations was very similar to what was previously reported even though rtL180M (29.1%) was found more often than rtM204V/I (26%)
. Even in previously treated or in currently treated patients, the rate of 3TC-resistance reached only 35%. These data could be explained by a low viral replication before treatment initiation, as low viral replication or an HBe-negative status was shown protective for resistance development in a previous study
. The cross-sectional nature of our study did not allow us to collect this information. This relatively low prevalence of resistance is incentive in considering the possibility of low level of compliance to anti-HBV drugs in our cohort.
Adefovir resistance mutations were exceptionally detected (rtA181T-change in 2.2%) and no change was observed for rtN236. Even if one considers that the rtA181T-strains have a slight decreased sensitivity to TDF, it is reassuring to observe a very low prevalence of ADV-resistance, leaving the opportunity to treat these patients with TDF
. The rtA181T or -V change is rarely selected by 3TC treatment and one can not exclude that this was not the case for these patients
[30, 31]. HAART often contains TDF, in addition to FTC or 3TC, a drug that has not been associated with selection of resistant strains so far
[32, 33]. The use of 3TC in association with TDF in patients from our cohort likely explains the low rate of resistance in patients receiving 3TC and strengthens the benefit of using potent antivirals such as TDF.
Consequence of drug-resistant variant selection on HBs reading frame has been extensively analyzed in our study. Contrary to what has been reported elsewhere, we did not observe significant changes on residues known to affect HBs "a" determinant
. In our hands, HBV genotype was the most influencing parameter to explain HBsAg variability and past or current treatment did not seem to particularly affect HBs antigenicity
. Key amino-acid changes known to particularly affect HBsAg detection by current immuno-assays or escape vaccination induced antibodies were found in less than 3% of patients
. Even though HBV vaccination is less immunogenic in HIV-infected individuals, the low prevalence of immune escape variants in this population should certainly be a strong argument for vaccination campaign to avoid further contamination in at-risk situations
The recognition of significant liver fibrosis in more than half of co-infected and HBV viremic patients in this series, along with a surprising high rate of suboptimal anti-HBV therapy use in this population, point out suboptimal clinical management of HBV infection in HIV-positive patients in France. Such observation was previously reported in several developed countries and highlights the need to promote and improve the education of physicians in charge of HIV-HBV infected individuals
Our study has some limitations due to the recruitment design. It is a transversal study that included patients on the basis of an HBV replication over 1,000 IU/mL whereas no information was collected on the reason for HBV virology assessment. Yet, these patients are clearly those who should benefit from an HBV-antiviral treatment and are those encountered daily in our practice. Noteworthy and contrary to what has been shown by several studies, no direct link was found between liver disease and HBV-VL
. This discrepancy might be explained by the relatively low number of patients with complete liver disease assessment in our study or by a different impact of HBV-VL in the context of HIV infection.
In conclusion, our study gives a new insight into the current epidemiology and clinical features of HBV-HIV infection in France. The strong link between the modes of contamination, the contamination event timeline by both viruses and the targeted population suggests that clinical management may be optimized according to this information. Even though potent active drugs with activities against both viruses are available, many patients have not received any anti-HBV treatment despite being treated by HAART and having for half of them advanced liver fibrosis. In addition, persistent HBV viremia in TDF treated patients suggests either non-compliance or suboptimal drug efficacy in this cohort. Finally, the low prevalence of circulating HBsAg immune-escape variants strengthens the need for an active promotion of anti-HBV vaccination of non-immune HIV-infected patients.