High Prevalence of Hepatitis C Virus Type 5 in Central France Evidenced by a Prospective Study from 1996 to 2002

Abstract

From 1996 to 2002, hepatitis C virus (HCV) typing was prospectively performed for 1,281 unselected HCV-infected and viremic patients, irrespective of their clinical status. Eighty-three patients (6.5%) were coinfected with human immunodeficiency virus (HIV) and HCV. A total of 1,195 strains were identified by a serotype screening (Murex HCV Serotyping 1-6 assay) and/or genotyping (Inno-LiPA HCV II) test. The distribution of HCV types showed an unusually high rate of type 5 (14.2%) that was stable over time and was the third most frequent type, after type 1 (59.1%) and type 3 (15.1%). HCV type 5 was more frequent in patients who were older than 50 (P = 10⁻⁶), but its frequency did not differ significantly by gender (P = 0.21). Serotyping was performed for 1,160 strains but failed for 30.2% of them. The efficiency depended on HIV status (for HCV-HIV-coinfected patients, half of the strains were untypeable) and HCV type. Genotyping was performed for 428 samples, with an overall efficiency of 99.3%. It failed in three cases, which were subsequently identified as HCV type 2. Serotyping and genotyping results for 39 patients showed discrepancies between the two methods for 4 patients, who had HCV type 2, type 6, or mixed infections. Thus, HCV type 5 may also be encountered frequently in Western countries. Its apparent confinement to a restricted area raises the question of how it emerged and underscores the need for further studies of HCV type prevalence, routes of transmission, pathogenicity, and responses to treatment.

Hepatitis C virus (HCV) is the main cause of parenteral non-A, non-B hepatitis, which leads to chronic liver disease and hepatocarcinoma. Since the first description of the HCV genome (5), numerous complete or partial nucleotide sequences of HCV isolates have been reported worldwide (5, 19, 22, 31). A comparison of these sequences revealed genetic heterogeneity throughout the viral genome, and a consensual international classification was made in 1994 (32). To date, six major genotypes and multiple subtypes have been identified. However, the study of viral diversity is still of crucial interest, as it is needed to elucidate the disputed role of different HCV genotypes in the progression and pathogenesis of liver disease. More information is needed on the epidemiological features and modes of HCV transmission in order to improve current molecular and diagnostic tools. This could be done by testing recent circulating isolates (28), which would greatly contribute to therapeutic strategies and the development of a vaccine (38).

The predominant genotypes are types 1, 2, and 3. They are distributed worldwide, but their relative frequencies differ greatly between countries (3). Genotypes 4 to 6 have been found mostly in less industrialized countries (India and countries in Southeast Asia) and the Middle East. Genotype 5 seems to be confined to South Africa (10, 38).

Recent reports have provided important new epidemiological data. HCV's high level of diversity was confirmed by more extensive studies of the type distribution in Africa and Southeast Asia, where, unlike in Europe and the United States, HCV has been endemic for a long time (24, 34). A spontaneous recombinant isolate was reported in 2002 from St. Petersburg, Russia (14). In Europe, there have been dynamic changes over time in the prevalence of the different HCV subtypes. For example, the prevalence of HCV 1b has decreased, while conversely, that of HCV 1a and 3a has increased (4, 6, 25, 26). Some rare genotypes have recently emerged in the Western world. Genotype 4 has been reported to exist in France, in the northern Parisian suburbs (20), in the south of Spain (27), and in Italy (18). In 2001, an unusually high prevalence of genotype 5 was reported for the region of Alicante in Spain (13).

In addition to its epidemiological relevance, viral typing might have a clinical impact. Numerous studies have reported a relationship between the HCV type and the response to interferon or pegylated interferon therapy, given alone or in combination with ribavirin (1). Patients infected with HCV type 2 or 3 have a better response to treatment than those infected with HCV type 1. Consequently, HCV type determination is now routinely performed when therapy is indicated.

Since 1996, we have prospectively performed viral typing at the laboratory of virology of the Clermont-Ferrand teaching hospital in Central France for all patients with chronic HCV infections. We used a two-step process involving a serotyping test and a genotyping test, and if necessary, sequencing of the 5′ untranslated region (UTR) and/or NS5 gene, and assessed its performance. Overall, 1,281 patients were included between 1996 and 2002 and 1,195 HCV typing results were obtained. They revealed an unusual local pattern of HCV distribution with a high prevalence of HCV type 5.

MATERIALS AND METHODS

Patients.

In a prospective analysis from 1996 to 2002, 1,281 consecutive patients were found to be chronically infected with HCV, and their HCV strains were characterized. The patients either had been admitted to the teaching hospital of Clermont-Ferrand (n = 986) or to nearby hospitals (n = 38) or were outpatients receiving medical care (n = 257). Ages were known for 1,258 (98.2%) of the 1,281 patients (median, 46.0 ± 15.9 years). For reasons of anonymity, gender was undisclosed for 33 (2.6%) patients: of the remaining patients, 530 (41.4%) were female and 718 (56%) were male. Eighty-three (6.5%) were coinfected with human immunodeficiency virus (HIV) and HCV.

Diagnosis of HCV infection.

We developed a cost-effective protocol in our laboratory for the diagnosis of HCV infection. Serum samples were tested for the presence of anti-HCV antibodies by two immunoassays, Axsym System HCV, version 3.0 (Abbott, Rungis, France), and Monolisa anti-HCV Plus, version 2 (Bio-Rad, Marnes la Coquette, France). When anti-HCV antibodies were detected in a given sample by both immunoassays, the presence of HCV genomic RNA was systematically tested by the use of a Cobas Amplicor HCV test, version 2.0 (Roche Diagnostics, Meylan, France). If the test was positive, the PCR products were properly divided into aliquots and stored at −80°C.

HCV identification with serotyping and/or genotyping tests and sequencing.

The identification of HCV strains was performed in the same reference laboratory throughout the study (see below).

For HCV RNA-positive patients, HCV identification was systematically carried out by serotyping with the Murex HCV Serotyping 1-6 assay (Abbott) for 1,160 of 1,281 (90.6%) consecutive patients. When the serotyping test failed, identification was performed by genotyping with the InnoLiPA HCV II assay (Bayer, Puteaux, France).

The genotyping of HCV strains is mandatory for their inclusion in multicentric therapeutic protocols. We performed genotyping tests directly for 121 (9.4%) of the 1,281 patients and after serotyping for 39 patients (3%). All of the procedures were performed and the results were interpreted according to the manufacturer's recommendations.

Nucleotide sequencing of the 5′ UTR and the NS5B gene was performed when HCV identification was not possible with the InnoLiPA HCV II genotyping test or when there were discrepancies between the serotyping and genotyping results. In addition, NS5B gene sequencing was also used to verify the identification of type 5 for randomly selected patients.

Sequencing of 5′ UTR.

Sequencing of the 5′ UTR was performed with the PCR products obtained with the Cobas Amplicor HCV test. A Big Dye Terminator, version 1.1, cycle sequencing kit (Applied Biosystems, Courtaboeuf, France) was used for direct sequencing of the purified DNA fragments. Sequencing reactions were performed in a final volume of 10 μl, including 1.5 μl of template PCR product, 4 μl of the sequencing mix, and 3.2 pmol of a primer described by Doglio et al. (8).

Sequencing of NS5B gene.

We adapted a previously described method (28) for sequencing of the NS5B gene. Briefly, a 401-bp DNA fragment was amplified by reverse transcription-PCR with primers PR1 and PR2, as described by Enomoto et al. (9). Sequencing of the purified DNA fragments was performed on both strands with the PR1 and PR2 oligonucleotide primers under the conditions described above for sequencing of the 5′ UTR.

For two patients, sequencing of the NS5B gene was kindly performed by Visible Genetics.

Sequence analysis and genotyping of HCV strains.

Electrophoresis of the purified sequencing products was performed with an ABI 310 genetic analyzer (Applied Biosystems). The sequences were then analyzed with the EMBL data library by comparisons with HCV prototype sequences, and the genotype was identified.

Statistical analysis.

All statistical analyses were performed with SPSS, version 10.0, software. The χ² test was used to compare categorical data, and a Kruskal-Wallis nonparametric test for variables with a non-Gaussian distribution was also used. The level of statistical significance was set at P values of <0.05.

Nucleotide sequence accession numbers.

The sequences of the NS5B region of 20 HCV type 5 strains were deposited under accession numbers AJ608776 to AJ608785, AJ626929, and AJ626990 to AJ626998.

RESULTS

Relative performances of serotyping and genotyping tests. (i) Serotyping of HCV strains from 1,160 patients.

We performed serotyping on serum samples from 1,160 patients. Eight hundred ten (69.8%) HCV strains were typed, 807 of which were single infections and 3 of which were from mixed infections (types 1 and 3, types 1 and 4, and types 3 and 4). The remaining 350 HCV strains (30.2%) were untypeable. Of the 350 failures, 196 (56%) failed due to an absence of reactivity and 154 (44%) failed because of nonspecific reactivity. The serotyping efficiency depended on the HCV type and the HIV serological status: 43 (51.8%) of the 83 strains from HIV-HCV-coinfected patients were untypeable compared to 307 (26.5%) of the 1,160 strains from HIV-negative patients. The serotyping test was more efficient for characterizing HCV type 1 than type 5 (Fig. 1).

FIG. 1.

Performance of serotyping test for identifying 1,073 strains according to HCV type. The identification of HCV strains was performed with the Murex HCV Serotyping 1-6 assay. Successful identifications are indicated in dark gray. Strains that were not identified by the serotyping test because of failure but were subsequently identified by the genotyping assay are indicated in white. Of 810 successful serotyping tests, 805 are presented, and the two infections with type 6 and three mixed infections have been excluded.

Of the 350 samples whose strains were untypeable, 268 (76.6%) were properly divided into aliquots and stored at −80°C to be used in a further genotyping test.

(ii) Genotyping of HCV strains from 428 patients.

A total of 428 samples underwent genotyping: they included the 268 serotyping failures and 160 samples from the patients included in the protocols, 121 of which were typed directly and 39 of which were retyped after the serotyping test (see Materials and Methods).

Using the Inno-LiPA HCV II test, we typed 425 of 428 (99.3%) strains, 424 of which were single infections and 1 of which was from a mixed infection (types 1a and 3a). The genotyping test failed for 3 (0.7%) of the 428 strains. All three of these had unidentifiable band patterns and were subsequently identified as type 2 by serotyping or sequencing of the 5′ UTR (type 2a).

(iii) Comparison of serotyping and genotyping for 39 patients.

For the 39 strains from patients who were included in protocols requiring a genotyping test, the HCV type was determined by both methods, thereby allowing a comparison of the two tests. Thirty-five unambiguous results were obtained. Discrepant results were observed for 4 (10.2%) of the 39 double tests. Two results corresponded to mixed infections, one with types 1 and 4 and one with types 3 and 4 (testing as types 4 and 3, respectively, by the genotyping test). For the two other cases, sequencing of both the 5′ UTR and the NS5B gene was performed. A serotype 6/genotype 2 strain was identified as being type 2 with two concordant sequencing results. This could be explained by a sequential infection, one in the past by a type 6 strain (detected by serotyping) and another more recent one by a type 2 strain (detected by genotyping). A serotype 6/genotype 1b strain was first identified as either type 1b (5′ UTR) or clade 6 (NS5B) and finally classified as type 6 because 5′ UTR sequencing discriminates poorly between types 1 and 6.

Overall distribution of HCV types in 1,195 patients.

Unambiguous results were obtained for 1,195 of 1,281 HCV-infected patients. Eighty-six (6.7%) types could not be determined either because the serotyping test failed or the samples had not been properly stored (n = 82), because there were discrepant results between the serotyping and genotyping assays (n = 2), or because the results obtained were not clear-cut (n = 2).

The distribution of HCV types among the 1,195 patients showed that type 1 was the most prevalent (n = 706 [59.1%]), followed by type 3 (n = 180 [15.1%]), type 5 (n = 170 [14.2%]), type 2 (n = 77 [6.4%]), and type 4 (n = 60 [5%]). Only one infection of type 6 (0.1%) and one unambiguous mixed infection with types 1 and 3 (0.1%) were observed (Fig. 2).

FIG. 2.

Distribution of HCV types among 1,195 patients attending Clermont-Ferrand teaching hospital from 1996 to 2002.

The year-by-year distribution has been stable since 1996 (the beginning of the prospective study), including the high percentage of type 5 strains (data not shown).

HCV type 5 identification by direct NS5B gene sequencing.

To confirm the unusual presence of HCV type 5, we randomly selected 20 HCV type 5-infected patients. Sequencing of the NS5B region of the corresponding strains confirmed the typing results obtained with the serotyping (n = 15) and genotyping (n = 5) assays.

HCV type distribution according to characteristics of non-type 5- and type 5-infected patients.

The characteristics of the 1,195 patients for whom typing of the HCV strain was successful were analyzed. Two patients, who were infected by type 6 and types 1 and 3, were excluded from the statistical analysis. An age was available for 1,173 (98.2%) of 1,195 patients (Fig. 3), and gender was available for 1,176 patients (98.4%).

FIG. 3.

Distribution of HCV types according to age among 1,173 patients. The types are indicated in the bars by different shades of gray, representing, from bottom to top, types 1, 2, 3, 4, and 5.

In non-type 5-infected patients, there was a significant variation in the distribution of types in the different age groups (P = 10⁻⁷, n = 1,173). Type 1 was the most frequent type in each age group, ranging from 56.1% (184 of 328) in the group of ≥60-year-olds to 62.2% (196 of 315) in the 40- to 49-year-old group. Type 2 was more frequent in the 50- to 59-year-old group (24 of 156 [15.4%]) than in the other groups (51 of 1,017 [5.0%]). Type 3 was more common in patients younger than 39 years old (93 of 374 [24.9%]) than in those older than 50 years old (24 of 484 [5%]). Type 4 was rare in the population studied, regardless of the age group. The only significant difference according to gender was observed for the distribution of HCV type 3 (for type 3 versus types 1, 2, 4, and 5, P = 0.005 and n = 1,176), which was found more frequently in men (121 of 179 [67.6%]) than in women (58 of 179 [32.4%]).

In type 5-infected patients, the virus was more frequent in patients older than 50 (125 of 484 [25.8%]) than in those younger than 49 (44 of 689 [6.4%]). The patients had a median age of 63.0 ± 16.5 years (range, 7 to 90 years), and 53.5% (85 of 159) were male. The proportion of patients older than 50 with type 5 infections was higher than the proportion of people older than 50 in the general population of the region (125 of 169 [74%] versus 493,026 of 1,308,656 [37.6%]; P = 10⁻⁶), but there was no significant difference in gender between the two (P = 0.21).

Preliminary epidemiological investigations of type 5-infected patients showed that the risk factor was either unknown or related to a history of blood transfusion.

DISCUSSION

The first aim of this prospective study was to analyze the distribution of HCV types in a general unselected population of 1,281 chronically HCV-infected patients seen in routine practice, irrespective of their clinical background. Thus, it can be reasonably assumed that any bias was negligible.

The main and surprising result of the study was the unusual distribution of HCV types in the area, with type 5 being the third most frequent. Figures from French reference centers show that the predominant genotypes in patients with chronic HCV infections are type 1 (58%), type 3 (22%), and type 2 (8 to 10%) (7, 17). Types 1 and 3 were also the two most prevalent types in our patients, with a similar rate for the first (59%) but a lower rate for genotype 3 (15.1%). However, for the first time in a Western country, there was a local pattern of HCV distribution in our study, with HCV type 5 accounting for 14.2% of the cohort. HCV genotype 5 has been observed in up to 30% of infections in South Africa (33, 35, 38), while small numbers of isolates have been identified in Canada (21), Belgium (36), and more recently, a region of Spain (13) and Brazil (16). In our study, infected patients were local inhabitants, and a preliminary investigation showed that they had not been at risk of acquiring HCV infection in other countries. Genotype 5 infection was more prevalent among individuals older than 50 years, including both men and women. Preliminary data indicate that genotype 5 was more frequent among patients with a history of blood transfusion, a finding that was also reported for Canadian blood donors (21) and Spanish patients (13) infected with HCV genotype 5. The fact that type 5 was seen in older patients from a largely settled semirural population strongly suggests that the type 5 infections were not recent but probably linked to some distant event, perhaps of iatrogenic origin. A retrospective analysis of HCV-positive samples stored in our laboratory before 1996 showed that type 5 has been present in our region since at least 1991 and that the number of type 5 infections was stable between 1996 and 2002 (data not shown). Hence, this unusual observation of type 5 is different from that of the emergence of HCV type 4 in the Parisian suburbs, which was probably related to the importation of viral isolates via African migrants (28). Throughout the study period, there was no local increase in the number of type 5 infections, and infection has not spread among the younger intravenous drug user community. This, together with the introduction of blood donor screening, suggests that there will be a reduction in genotype 5 infections in our region in the future. An epidemiological study and a phylogenetic analysis of these HCV isolates are currently in progress to investigate the source(s) of HCV type 5 infections and its mode of transmission.

Such a large-scale study was possible because we used an easy-to-perform process that was not too time-consuming or expensive. This process comprised the wide use of serological typing as a first step followed by the use of genotyping methods. The second aim of the study was to evaluate the performance and reliability of the methods performed in our routine practice.

The reference method for HCV genotype determination is sequence analysis followed by phylogenetic analysis for a definitive classification. Such a procedure is expensive, time-consuming, and not suitable for routine use. In addition, discrimination between subtypes is not necessary for the management of hepatitis treatment. Thus, an automatable serological typing assay was performed as a first step, followed, in the event of failure, by a standardized genotyping assay. The sensitivity of the serological typing assay was 70% in relation to the results obtained by the genotyping assay. This rate, which is lower than those previously reported (23, 37, 39), could be explained by the high prevalence of genotype 5 in the population studied. Gault et al. recently reported that a kit from the same manufacturer using a new version of the HCV serotyping 1-6 assay gave better results for rare types (11). A second possible explanation is that the serotyping tests were not sensitive enough for HIV-HCV-coinfected patients, since for half of them, the HCV strains were untypeable (11).

The sensitivity of the Inno-LiPA genotyping test was 99.5%. Overall, this performance was consistent with those described elsewhere (2, 12, 15, 39). The genotyping assay failed for three samples, which were later identified as type 2 by the serological method.

Of the four discrepant typing results, two were mixed infections identified by the serotyping test but not by the genotyping assay. While the presence of cross-reactivities cannot be excluded as a cause of mistyping by the serological assay, it is generally admitted that serological methods are more sensitive for detecting mixed or sequential infections (29, 30, 39). Molecular biology-based methods are less effective for identifying mixed infections because one strain is often predominant and is the only one amplified by PCR.

Several virus-typing methods based on amplified sequences from the 5′ UTR have been developed and present an advantage because reverse transcription-PCR of this region is routinely used for the diagnosis and follow-up of HCV infections. However, there are only a few sequence differences between HCV types in the 5′ UTR, and the sequences of different subtypes may even be identical (33). Hence, the NS5 region, which is more variable, would seem to be more suitable for HCV type discrimination, in particular for types 4, 5, and 6, which are less frequently encountered in the Western world (10). This discriminatory power was illustrated by the mistyping of one strain, which was identified as type 1b by 5′ UTR sequencing but as type 6 (or included in clade 6) by NS5B sequencing and the serological test.

Epidemiological surveys of the circulation of HCV strains and their evolution could be of great use for improving biological tests for diagnosis and follow-up and for developing specific prevention procedures. The HCV genotype distribution has a direct impact on medical practice and treatment. To date, however, no data have been available on the natural history of chronic hepatitis and the response to treatment of HCV genotype 5 infections. Characterization of the numerous type 5 infections and corresponding strains in our laboratory should greatly contribute to progress in these areas.