 |
|
|
| |
Prevalence of hepatitis C in an ethnically diverse HIV-1-infected cohort in south London: 25% of coinfected unaware they have HCV
|
| |
| |
"....The prevalence of HIV/HCV co-infection in this ethnically diverse HIV-1-infected cohort in south London was 8.9%, which is more than 100-fold higher than the HCV seroprevalence reported in the general UK population (0.4%).. may, however, be an underestimate of the true HCV prevalence... Eighty-two per cent of our HIV-infected patients with a current or previous history of IDU were HCV infected... these estimates highlight an urgent need to undertake a proactive programme of voluntary testing and counselling for HIV and HCV infection among high-risk groups, particularly among IDUs, to enable intervention and reduction of onward transmission. This should also include hepatitis A and B screening and vaccination when appropriate... As it is now proposed that all gay and bisexual men attending genitourinary clinics should have been offered HBV immunization at their first visit, this presents an additional opportunity for HCV screening.."
HIV Medicine
May 2005
AH Mohsen, S Murad and PJ EasterbrookDepartment of HIV/GU Medicine, The Guys Kings' and St Thomas School of Medicine, Kings' College Hospital, London, UK
ABSTRACT
Objectives: There is limited information on the prevalence of and risk factors for hepatitis C virus (HCV) infection among HIV-1-infected patients in the UK. Our objective was to determine the prevalence of HCV infection among an ethnically diverse cohort of HIV-infected patients in south London, and to extrapolate from these data the number of co-infected patients in the UK.
Methods: A total of 1017 HIV-1-infected patients who had attended King's College Hospital HIV clinic between September 2000 and August 2002 were screened for HCV antibody using a commercial enzyme-linked immunosorbent assay (ELISA). Positive results were confirmed by polymerase chain reaction (PCR) or recombinant immunoblot assay. Demographic, clinical and laboratory data were obtained from the local computerized database and medical records. We applied our HCV prevalence rates in the different HIV transmission groups to the estimated number of HIV-infected persons in these groups in the UK, to obtain a national estimate of the level of HIV-HCV co-infection.
Results:
Of the 1017 HIV-1-infected patients, 407 (40%) were white men, 158 (15.5%) were black African men, 268 (26.3%) were black African women, and 61 (6%) and 26 (2.6%) were black Caribbean men and women, respectively.
Heterosexual exposure was the most common route of HIV acquisition (53.5%), followed by men having sex with men (36.9%), and current or previous injecting drug use (IDU) (7.2%).
The overall prevalence of HCV co-infection was 90/1017 (8.9%), but this varied substantially according to route of transmission, from 82.2% among those with a history of IDU (which accounted for 67% of all HCV infections), to 31.8% in those who had received blood products, to 3.5% and 1.8% in those with homosexually and heterosexually acquired infection, respectively.
Multivariate logistic regression analysis identified several independent risk factors for HCV infection: a history of IDU [odds ratio (OR)=107.2; 95% confidence interval (CI)=38.5-298.4], having received blood products (OR=16.5; 95% CI=5.1-53.7), and either being from a white ethnic group (OR=4.3; 95% CI=1.5-12.0) or being born in Southern Europe (OR=6.7; 95% CI=1.5-30.7).
Based on the 35,473 known HIV-1-infected persons in the UK and the 10,997 estimated to be unaware of their status, we projected that there are at least 4136 HIV-HCV co-infected individuals in the UK and 979 who are unaware of their status.
Conclusions: Overall, 9% of our cohort was HIV-HCV co-infected. The prevalence was highest among intravenous drug users (82%), who accounted for most of our HCV cases, and lowest among heterosexual men and women from sub-Saharan Africa and the Caribbean [< 2%]. Our estimate that a significant number of co-infected persons may be unaware of their HIV and HCV status, highlights an urgent need to increase the uptake of HCV and HIV testing, particularly among injecting drug users, to reduce the risk of onward transmission.
AUTHOR DISCUSSION
The prevalence of HIV/HCV co-infection in this ethnically diverse HIV-1-infected cohort in south London was 8.9%, which is more than 100-fold higher than the HCV seroprevalence reported in the general UK population (0.4% [35]). However, this is considerably lower than the overall HCV prevalence in various HIV-infected European (16.8 to 40%) [2,10-15] and North American (14.4 to 55%) [16-23] cohorts, and mainly reflects our low proportion (<10%) of injecting drug users (IDUs) and those infected via blood products, as well as our higher proportion of patients from sub-Saharan Africa with heterosexually acquired infection, who had the lowest HCV prevalence. Our seroprevalence rates may, however, be an underestimate of the true HCV prevalence because we tested only a single sample early in the course of infection, and used only an ELISA assay. At least 4% of HIV/HCV co-infected patients have no detectable antibodies in the presence of HCV viraemia [38,39], or may lose detectable antibodies from serum despite persistent viraemia as a result of immunosuppression [40].
When we stratified our HCV prevalence rates according to HIV transmission group, there was greater consistency with the rates reported from other studies. Eighty-two per cent of our HIV-infected patients with a current or previous history of IDU were HCV infected, which is consistent with previously reported data of over 80% in five studies [12,18,19,22,23,26,41], including two from Southern Europe comprising over 6000 patients [12,41]. Of note, the majority of our IDUs were white (86.3%), and were born in either the UK (39.7%) or Southern Europe (34.3%), mainly Italy (83%), with the remainder from Spain, Portugal and Greece. IDU was also a very strong independent risk factor for HCV infection (OR=135.5; 95% CI 47-389), consistent with all previous studies. We found that homosexual or bisexual men with a history of IDU were less likely to be HCV positive than heterosexuals with a similar history (55.5 vs. 85.9%), which is consistent with data from the USA [18,20]. This is likely to be explained by the fewer high-risk drug use practices such as needle sharing among our homosexual patients, although drug use behaviours were not assessed in our study.
Of the 22 patients with a history of receiving blood products (only 2.2% of our cohort), seven patients (31.8%) were co-infected with HCV, and as expected this was a strong independent risk factor for HCV positivity (OR=18.1; 95% CI 5.7-57.5). Of note, none of the patients were haemophiliac, and they had either been exposed in sub-Saharan Africa, where blood product screening has been less systematic, or were exposed before the introduction of screening in Western Europe.
The overall prevalence of HCV in our homosexual/bisexual patients was 3.5%, which is consistent with the rates reported from most other studies of between 2.6 and 9% [4,12,19,27]. The HCV prevalence among those with heterosexually acquired infection was 1.8%, which is lower than most of the rates reported previously (6% [12], 9.9% [4] and 25.6% [21]), and is likely to be a result of the high proportion of our heterosexual risk group who were from sub-Saharan Africa, for whom the prevalence was lower (1.9%) than for white patients in this group (5%).
There were some important differences in HCV prevalence amongst different ethnic groups. The prevalence among white men and women was 16%, compared with 2.4% for black Africans and 1.1% for black Caribbeans. In the multivariate analysis, white ethnic group and being born in Southern Europe were associated with an increased risk of HCV infection, even after adjustment for IDU, although the odds ratio was substantially reduced from 8-fold in the univariate analysis to 4-fold in the multivariate analysis for white race, and from 93- to 7-fold for being born in Southern Europe. Of note, all but one of the 25 co-infected patients born in Southern Europe were IDUs. The finding of a borderline association of being born in South America and HCV infection was not explained by drug use. Of the 14 co-infected patients, six were homosexual/bisexual men from Brazil and seven were heterosexual, mainly from Brazil, but also from Columbia and Ecuador. This contrasts with data from the USA, in which a higher HCV prevalence was found among black Americans (56.3%) and Hispanics (17%) compared with white Americans (24.7%), but importantly IDU was more frequent in these other ethnic groups [20].
We obtained a crude estimate of the likely numbers of co-infected patients in the UK based on local co-infection rates in different risk groups in our south London HIV-infected patient cohort, which were then applied to the number of known as well as estimated number of unknown HIV infections in the UK. We estimated that there are 4136 HIV/HCV co-infected patients in the UK (95% CI 3207-5066) and the majority of these (n=2285; 55.3%; 95% CI 1988-2508) are current or former IDUs. Overall, about 979 (95% CI 759-1199) are unaware of their HIV infection and probably also their HCV infection. The greatest concern in terms of onward risk of HIV and HCV transmission is the estimated 412 (95% CI 358-452) co-infected IDUs in the UK who are unaware of their status. However, no information is available on the proportion of these individuals who maintain high-risk drug use practices, such as sharing of needles or other equipment [42]. Although one caveat in the interpretation of these estimates is that our HIV cohort in south London is not representative in risk group, gender and ethnic distribution of HIV-infected patients elsewhere in the UK, it is unlikely that the prevalence of co-infection within the different risk groups would vary substantially by region.
Overall, these estimates highlight an urgent need to undertake a proactive programme of voluntary testing and counselling for HIV and HCV infection among high-risk groups, particularly among IDUs, to enable intervention and reduction of onward transmission. This should also include hepatitis A and B screening and vaccination when appropriate, although a proportion of our IDUs are already infected with the hepatitis B virus (HBV). Opportunities for screening include drug treatment centres, drop-in clinics, general practice surgeries, prison and remand centres, mental health services, and hospital wards and out-patient clinics. Although the prevalence among our gay men was relatively low (3.5%), this is still at least 7-10-fold higher than in the general population. As it is now proposed that all gay and bisexual men attending genitourinary clinics should have been offered HBV immunization at their first visit, this presents an additional opportunity for HCV screening [43].
A further priority is to generate more accurate estimates of the local, regional and national burden of severe chronic HCV infection in HIV-infected persons in the UK. Basic models have been developed for HCV monoinfection [44], but these require further modification to take into account the higher rate of progression among co-infected patients, non-HIV- but HCV-related mortality, and a rising incidence of cirrhosis and hepatocellular carcinoma, as the majority of co-infected patients acquired their HCV infection before their HIV infection, often more than two decades ago. On the basis of such models, it will then be possible to estimate treatment costs, which are likely to be considerable given the frequent recognition of fibrosis in HCV-infected patients who undergo liver biopsy [45]. Although currently available HCV therapy with pegylated interferon and ribavirin is expensive, a recent cost-effectiveness analysis concluded that, even with mild fibrosis, early therapy is cost-effective [22]. In addition, the response rate in nongenotype-1 HCV co-infected patients is comparable to that observed in patients not infected with HIV [46,47].
INTRODUCTION
Co-infection with hepatitis C virus (HCV) in HIV-infected patients has several important implications for clinical care and management. First, the rate of progression of hepatitis C to end-stage liver disease is accelerated 3-fold in the presence of HIV infection, and HCV infection is now a leading cause of hospitalization and death in this population [1]. Secondly, hepatitis C increases the risk of hepatotoxicity with antiretroviral therapy, particularly with drugs such as nevirapine, and so may impact on the timing and choice of therapy [1,2]. Thirdly, the use of highly active antiretroviral therapy (HAART) in HCV-infected patients may precipitate an HCV immune reconstitution phenomenon, and there may also be impaired immune recovery in co-infected individuals [3,4], although the data remain equivocal. Finally and most importantly, there is now effective therapy available for the treatment of hepatitis C in both monoinfected and to a lesser extent HCV/HIV co-infected patients, with a sustained virological response rate of 40% in HCV genotypes 2 or 3 [5-7]. In recognition of the importance of HCV as a pathogen in HIV-infected individuals, it is now recommended that all HIV-infected individuals be screened for HCV infection [8,9].
Because the prevalence of hepatitis C in HIV-infected populations varies widely according to the route of acquisition [from 3-15% in homosexual/bisexual men, to 80% in injecting drug users (IDUs), to 98% in haemophiliacs], the prevalence of hepatitis in different HIV-infected cohorts is largely determined by the proportion of the different risk groups represented. As a result, seroprevalence rates vary considerably, from 17 to 40% across Western Europe [2,10-15] and from 14 to 55% across the USA [16-23]. Table 1 summarizes the HCV seroprevalence rates in 25 cohorts worldwide: 90% of HIV+ IDUs have HCV in the USA, in Europe & other areas of the World 50-90%, China 99%, Braz.
Although it is estimated that there are between 200 000 and 400 000 people infected with HCV in the UK [34,35], the number co-infected with both HCV and HIV is unknown. This paper describes the prevalence of HCV in an ethnically diverse HIV-1-infected cohort in south London. We also applied our HCV prevalence rates in the different transmission risk groups to the total number of persons estimated to be HIV-infected in these groups in the UK, in order to estimate the magnitude of co-infection in the UK HIV-1-infected population.
RESULTS
Characteristics of the study population
Overall, 407 patients (40%) in our cohort of 1017 HIV-1-infected patients were white men, 45 (4.4%) were white women, 158 (15.5%) were black African men, 268 (26.3%) were black African women, and 61 (6%) and 26 (2.6%) were black Caribbean men and women, respectively. Heterosexual exposure was the most common HIV transmission group (53.5%), followed by men having sex with men (36.9%). Only 7.2% of patients gave a history of current or previous IDU as their main risk for HIV acquisition. The largest group was born in sub-Saharan Africa (40.5%) [south-east Africa (n=155), East Africa (n=150) or West Africa (n=107)], followed by the UK (38.4%, n=389), other parts of Western Europe (6.3%, n=65), the Caribbean (5.6%, n=57), Southern Europe (3.4%, n=35) and Northern Europe (0.3%, n=3). Only 11.8% were born outside these areas. At first HIV presentation, the median age was 32.6 years [interquartile range (IQR) 28.5-37.7 years], and the median CD4 cell count and viral load were 288 cells/muL (IQR 134-442 cells/muL) and 15 884 HIV-1 RNA copies/mL (IQR 2093-69023 copies/mL), respectively.
Prevalence of hepatitis C
The overall prevalence of hepatitis C was 8.9% (90 of 1017 patients). The prevalence varied considerably according to HIV transmission risk group, ranging between 1.8% in those with heterosexual exposure to 82% in patients reporting a history of IDU.
Table 2 compares the distribution of characteristics in the 90 HIV/HCV co-infected and 927 non-HCV-infected patients. There were no statistically significant differences between the groups in age (33 vs. 32.6 years, respectively; P=0.54), CD4 cell count (282 vs. 288 cells/muL; P=0.78), or viral load at the first presentation (12 000 vs. 16 505 copies/mL; P=0.34) or gender. However, patients infected with HCV were more likely than non-HCV-infected patients to be IDUs (66.7% vs. 1.4%, respectively), more likely to have a history of receiving blood products (7.8% vs. 1.6%), and less likely to be homosexual (14.4% vs. 39.1%) or heterosexual (11.1% vs. 57.6%) (overall chi2P<0.001). In addition, they were more likely to be white (81.1% vs. 40.9%, respectively), and conversely less likely to be nonwhite [black African and black Caribbean combined; 12.2% vs. 54.1%; overall chi2P<0.001]. They were more likely to be born in Southern Europe (28.1% vs. 1.1%, respectively), and less likely to be born in sub-Saharan Africa or the Caribbean basin [13.5% vs. 50.1%; overall chi2P<0.001].
Risk factors for HCV infection in HIV-positive patients
In a multivariate logistic regression analysis, we examined three variables that had shown a significant association in univariate analyses with HCV infection: risk group, ethnic origin and place of birth. Because ethnic origin and country of birth are highly correlated, we analysed these variables in two separate models. A history of IDU [odds ratio (OR) 107.2; 95% CI 38.5-298.4] and having received blood or blood products (OR 16.5; 95% CI 5.1-53.7) were independently associated with HCV co-infection. In one analysis, which included ethnic group as a variable (model 1), white race was also associated with an increased risk of HCV infection (OR 4.3; 95% CI 1.5-12.0). In the analysis with place of birth instead of ethnic group in the model (model 2), being born in Southern Europe (OR 6.7; 95% CI 1.5-30.7) or South America (OR 6.7; 95% CI 1.1-41.2) was also associated with an increased risk of HCV infection.
Estimate of total number of HIV/HCV co-infected patients in the UK
We performed a crude estimation of the total number of HIV/HCV co-infected patients in the UK, based on the prevalence data from our local south London cohort. Table 4 shows the total number of people living with HIV in the UK up to June 2003 in the three main transmission risk group strata (column 2). The additional number who were HIV-1 infected but unaware of their status (column 4) was estimated by applying the percentage with unknown HIV status in homosexual men (22%), IDUs (22%) and heterosexuals (43%), based on UK data from the UK Unlinked Anonymous Seroprevalence Survey [36] (column 3) at the end of 2001, to the total numbers in each transmission risk group strata in the UK (column 2). We estimated that there are approximately 4136 HIV/HCV co-infected patients in the UK, of whom 2285 (55.3%) are IDUs, 711 (17.2%) are homosexual and 356 (8.6%) are heterosexual (column 7). We also estimate that 412, 128 and 107, respectively, in these groups are unaware of their HIV and probably HCV status (column 6). Because the UK national data includes a proportion of individuals with missing risk factor information, and those with mother-to-child transmission for which we have no hepatitis C prevalence rates, we were not able to estimate the rate of co-infection in these transmission groups.
Patients and methods
Study population and data collection
This cross-sectional seroprevalence survey was based on all HIV-1-infected patients (n=1017) who had attended the Kings College Hospital HIV clinic at least once between 1 September 2000 and 31 August 2002. A total of 773 patients (76%) had already been tested for HCV as part of the serological screen that is performed routinely at their initial clinic visit. In the remaining 244 (24%) patients, a serum sample from their initial or an early visit that had been stored at -70 °C was tested retrospectively. This test was conducted anonymously using the clinic ID number with the approval of the local ethics committee. All these patients have subsequently been approached for consent to test through the routine clinical service. Epidemiological data were abstracted from the clinic computerized database or medical records on all patients, and included: date and age at initial HIV diagnosis, gender, ethnic origin, country of birth, HIV risk group, and CD4 cell count and viral load at first clinic visit.
Laboratory methods
HIV-1 infection was defined by positivity on a fourth-generation enzyme-linked immunosorbent assay (ELISA) (Genescreen Plus HIV Ag-AB; Bio-Rad Laboratories, Richmond, CA), and viral detection of HIV RNA (Roche Amplicor version 1.5; Roche Diagnostic, Basel, Switzerland). HCV serology was performed using a third-generation ELISA (ABBOTT HCV version 3.0; Abbott Laboratories, Abbott Park, IL) in 775 patients and using ORTHO HCV 3.0 (Ortho-Clinical Diagnostics, Raritan, NJ) in 244 patients. Positive results were confirmed by polymerase chain reaction (PCR) using either the Cobas Amplicor assay (version 2.0; Roche Diagnostic) or a recombinant immunoblot assay (RIBA HCV 3.0; Chiron Corporation, Emeryville, CA).
Statistical analysis
We examined the prevalence of HCV infection according to selected demographic and laboratory characteristics. These characteristics were also compared between the HIV/HCV co-infected patients using a chi2 or Fisher's exact test, or proportion tests for categorical variables, and a nonparametric two-sample median test for continuous data. P-values for multiple comparisons were Bonferroni adjusted. Univariate and multivariate logistic regression was used to identify independent risk factors for HCV infection. All 95% confidence intervals (CIs) in Table 4 (see later) were estimated by assuming binomial distributions, but for the overall prevalence a normal distribution was assumed. Data were analysed using stata (Version 6; STATA Corporation, College Station, TX), and all the P-values stated are two-sided.
The number of known persons currently living with HIV in the UK up to June 2003 [36], and the percentage estimated to be HIV-infected but unaware of their status in the three main risk group strata (sex between men, IDU and sex between men and women), based on data from the Unlinked Anonymous Seroprevalence Survey at the end of 2001, were obtained [37]. We estimated the prevalence of HCV in the UK HIV-infected population by applying the prevalence data with 95% CIs from our cohort, according to the various transmission risk group strata, to the national data.
|
|
| |
| |
|
|
|
|
|
