Patients
All HBV-infected Swiss residents who had HBV genotyping as part of their regular medical care performed in our reference laboratory between October 2002 and October 2013 were included in our retrospective study. A medical chart review was conducted to collect detailed information on the most important demographic and clinical characteristics. Demographic characteristics included age, sex, and region of origin. The region of origin was coded according to the patient’s nationality. In individuals who immigrated to Switzerland and were naturalized, the region of birth was considered. The study was approved by the local ethics committee of the University of Bern. Patient records were de-identified prior to analysis. In accordance with Swiss law and the Declaration of Helsinki no written informed consent was obtained for this retrospective study.
HBV genotyping
From 2002 to 2009 the HBV genotype was determined using the INNO-LiPA HBV Genotyping assay (Innogenetics N.V., Ghent, Belgium) according to the manufacturer’s protocol. From 2010 to 2013 HBV genotyping was performed by direct sequencing using primers published by Mallory et al. and Schildgen et al. [10, 11]. For both methods (INNO-LiPA HBV Genotyping assay and direct sequencing) viral DNA was extracted from the patient’s serum or plasma using NucliSENS easyMAG (bioMérieux, Paris, France) according to the manufacturer’s protocol. If genotyping was performed by direct sequencing, a fragment of 941 nucleotides of the viral polymerase/HBsAg was amplified in a primary PCR (pPCR) using the previously described primers HBV_1F and HBV_4R by Mallory et al. [10]. If needed, a nested PCR (nPCR) was performed using the primers TGGATGTGTCTGCGGC (sense primer) and CKTTGACADACTTTCCAATCAATAG (antisense primer) published by Schildgen et al. yielding a PCR product of 622 nucleotides [11].
All PCR products were analyzed by electrophoresis in a 1 % agar gel and then purified using QIAquick PCR Purification Kit (QIAGEN GMBH, Hilden, Germany) according to the manufacturer’s protocol. The purified amplicons were subjected to bidirectional Sanger sequencing using the previously described primer HBV_1F and the antisense primer published by Schildgen et al. [10, 11] for pPCR products and the primers published by Schildgen et al. [11] for nPCR products yielding a sequencing product of 815 nucleotides and 622 nucleotides respectively. Cycle sequencing was performed according to Platt et al. [12]. After purification of the cycle sequencing products by the QIAGEN DyeEx 2.0 Spin Kit (QIAGEN GMBH, Hilden, Germany) the electropherographs were acquired on a genetic analyzer AB-3130 (Life Technologies Europe BV, Nieuwerkerk, Netherlands) and then processed using SeqMan (DNASTAR Inc., Madison, WI, USA). For in silico sequence analysis and genotype determination the open access interpretation tool geno2pheno was used provided by Genafor at http://www.genafor.org/hbv/hbvpredict.php described by Beggel et al. [13]. By performing direct sequencing combined with in silico analysis using geno2pheno additionally to the genotype the subgenotype could be determined while by the INNO-LiPA assay the genotype only was identified.
Serologic analysis
If performed in our laboratory, samples tested before April 2009 were analyzed by Abbott AxSYM (Abbott Laboratories, Chicago, USA) for Anti-HBe, Anti-HCV and the HIV status. HBeAg was tested using VIDAS (bioMérieux, Paris, France). After April 2009 these serological parameters were tested using ARCHITECT i2000sr (Abbott Laboratories, Chicago, USA). Anti-HDV was tested with ETI-AB-DELTAK-2 (DiaSorin, Saluggia, Italy). All tests were performed according to the manufacturer’s instruction. If no HBeAg test, HIV, HCV or HDV serology was carried out in our institution, we reviewed the charts to identify analyses which were performed elsewhere.
Phylogenetic analysis
The phylogenetic analysis was performed using the Phylogeny.fr platform [14] with a set of 136 sequences (all samples from 2010 to 2013) from our study of HBV infected patients in Swiss residents [GenBank accession numbers KM524121 to KM524256] and compared to 15 reference strains from GenBank representing all genotypes from A to G. The sequences analysed were all restricted to the same 413 nucleotide stretch of the viral polymerase containing no gaps, deletions or insertions. The alignment was constructed using clustalW2 software provided by the European Bioinformatic Institute at http://www.ebi.ac.uk/Tools/msa/clustalw2/. The phylogenetic tree was calculated by an improved neighbour-joining method implemented in the BioNJ program [15] as part of the Phylogeny.fr platform. Bootstrap analysis was carried out 1000 times for reliability confirmation of the resulting structure. The phylogenetic tree was rooted for non-D genotypes.
Statistical analysis
In order to analyze the genotype distribution among patients living in Switzerland with different migration background, we classified all patients into one of five “regions of origin”: Switzerland, Europe/Mediterranean, sub-Sahara Africa, Asia and unknown/other. The distribution of regions of origin of our study population was compared with estimates from the general population living in Switzerland, obtained from the federal office of statistics [16]. Baseline characteristics of patients were compared across their original region of origin using Fisher’s exact test and ANOVA for categorical and continuous variables, respectively. The genotype distribution according to region of origin was described using proportional bar charts.
Risk factors for HBeAg-positivity were assessed using logistic regression. The following explanatory variables were included in univariable models: sex, age, region of origin, HBV genotype, HCV-coinfection, HIV-coinfection and serological evidence of HDV-infection. Variables significantly associated with the outcome variable (p < 0.05) were included in a multivariable regression model. Multiple imputation was used to impute missing HIV and HCV measurements at baseline, with analyses run on each of 20 datasets and results combined with Rubin’s rules [17]. In a sensitivity analysis, the results of multivariable analyses from the model using multiple imputation were compared to those obtained from complete-case analyses. All analyses were performed using Stata software version 12.0 (College Station, Texas, USA).