Hepatitis B or hepatitis C coinfection in HIV-infected pregnant women in Europe
Dr Claire Thorne, MRC Centre of Epidemiology for Child Health, UCL Institute of Child Health, University College London, 30 Guilford Street, London WC1N 1EH, UK. Tel: 44 207905 2105; fax: +44 207813 8145; e-mail: email@example.com
The aim of the study was to investigate the prevalence of and risk factors for hepatitis C or B virus (HCV or HBV) coinfection among HIV-infected pregnant women, and to investigate their immunological and virological characteristics and antiretroviral therapy use.
Information on HBV surface antigen (HBsAg) positivity and HCV antibody (anti-HCV) was collected retrospectively from the antenatal records of HIV-infected women enrolled in the European Collaborative Study and linked to prospectively collected data.
Of 1050 women, 4.9% [95% confidence interval (CI) 3.6–6.3] were HBsAg positive and 12.3% (95% CI 10.4–14.4) had anti-HCV antibody. Women with an injecting drug use(r) (IDU) history had the highest HCV-seropositivity prevalence (28%; 95% CI 22.8–35.7). Risk factors for HCV seropositivity included IDU history [adjusted odds ratio (AOR) 2.92; 95% CI 1.86–4.58], age (for ≥35 years vs. <25 years, AOR 3.45; 95% CI 1.66–7.20) and HBsAg carriage (AOR 5.80; 95% CI 2.78–12.1). HBsAg positivity was associated with African origin (AOR 2.74; 95% CI 1.20–6.26) and HCV seropositivity (AOR 6.44; 95% CI 3.08–13.5). Highly active antiretroviral therapy (HAART) use was less likely in HIV/HCV-seropositive than in HIV-monoinfected women (AOR 0.34; 95% CI 0.20–0.58). HCV seropositivity was associated with a higher adjusted HIV RNA level (+0.28log10 HIV-1 RNA copies/mL vs. HIV-monoinfected women; P=0.03). HIV/HCV-seropositive women were twice as likely to have detectable HIV in the third trimester/delivery as HIV-monoinfected women (AOR 1.95; P=0.049).
Although HCV serostatus impacted on HAART use, the association between HCV seropositivity and uncontrolled HIV viraemia in late pregnancy was independent of HAART.
Globally an estimated 350–400 million people are chronically infected with hepatitis B virus (HBV), 190 million are chronically infected with hepatitis C virus (HCV) and 33 million are living with HIV infection today [1,2]. As a result of shared routes of transmission, the HIV, HBV and HCV epidemics overlap, with around 10% of the HIV-infected population estimated to have chronic HBV infection and around a third estimated to have chronic HCV infection . Coinfection with these viruses is a growing problem and liver disease is a major cause of morbidity and mortality in HIV-infected people in resource-rich settings [3–5].
The clinical management of HIV-infected individuals coinfected with HCV or HBV is challenging, and HIV infection is known to have a negative impact on the outcome of HCV and HBV infections [6,7]. Although whether or not HCV directly impacts upon HIV disease progression remains controversial [3,8–10], the complex interactions between HIV/HCV coinfection and highly active antiretroviral therapy (HAART) use and the indirect effect of these on HIV disease progression are increasingly apparent [6,7,11]. Few studies have addressed the issue of coinfection with HCV and/or HBV in HIV-infected pregnant women to date [12–15] and, in particular, there are no data on the prevalence of HIV/HCV or HIV/HBV coinfection in antenatal populations in Europe. Understanding the epidemiology of HBV, HCV and HIV coinfection in pregnant women is important because of vertical transmission risks [16,17] and to inform clinical management.
We conducted a substudy within the European Collaborative Study (ECS) to estimate the prevalence of and risk factors for HCV and HBV coinfection [assessed as HCV antibody (anti-HCV) and hepatitis B surface antigen (HBsAg), respectively] in a population of HIV-infected pregnant women, and to compare HIV-infected women with and without markers for HCV and/or HBV coinfection with respect to immunological and virological characteristics and antiretroviral therapy (ART) use.
The ECS is a prospective cohort study in which HIV-infected pregnant women are enrolled and followed in pregnancy, and their children followed from birth, according to standard clinical and laboratory protocols . The cohort was initiated in 1985 and includes patients from centres in Western and Central Europe and, since 2000, Ukraine. At each centre, women identified as HIV-infected during pregnancy, as well as women identified as HIV-infected before pregnancy, were invited to participate. Informed consent and ethical approval were obtained according to local guidelines. Maternal information routinely collected included sociodemographic characteristics, obstetrical history and HIV-specific information, including ART, but HBV or HCV infection status was not routinely recorded. Laboratory variables collected included maternal CD4 cell counts and plasma HIV RNA viral loads. As a result of limited laboratory capacity at the time of the study, only 110 of 520 women (21%) from the Ukraine centres had baseline CD4 cell counts available and none had HIV RNA levels measured.
This nested study included women from 15 ECS centres in seven countries (Spain, Italy, UK, Belgium, Sweden, Germany and Ukraine) who delivered between January 1999 and October 2005 in all but the Ukrainian centres. This period was truncated to between January 2003 and October 2005 for the Ukrainian centres because of the large numbers of pregnant women enrolling . Information on markers of HBV and/or HCV coinfection (HBsAg and anti-HCV antibody) was collected retrospectively from the antenatal records of the included women. All centres screened all HIV-infected pregnant women for anti-HCV antibodies and HBsAg. HCV antibody testing was performed with commercial enzyme-linked immunosorbent assays and women were classified as HCV-infected based on a seropositive test. All analyses relating to HCV coinfection were based on HCV seropositivity only, and thus may have included some women with resolved HCV infection. Similarly, HBV coinfection was based on HBsAg positivity, leading to the inclusion of active infections, chronic infections and carriers. Classification of undetectable HIV viral load (viral suppression) was based on the assay lower limit of quantification. HAART was defined as a regimen of at least three antiretroviral (ARV) drugs including either a nonnucleoside reverse transcriptase inhibitor or a protease inhibitor (PI).
Univariable comparisons were assessed with the χ2 test for categorical variables. Logistic regression was used to obtain unadjusted and adjusted odds ratios (AORs) and 95% confidence intervals (CIs). The unadjusted and adjusted logistic regression analyses for HCV seropositivity were both limited to women with data available on all the following risk factors: maternal age, area of birth, injecting drug use(r) (IDU) history and HBsAg carriage; 125 women were excluded because of missing sociodemographic data. The risk factor analyses for HBsAg positivity included maternal age, area of birth, IDU history, history of sex with an IDU and HCV seropositivity; 137 women were excluded because of missing sociodemographic data. The logistic regression models for receipt of HAART were limited to women from Western European centres, reflecting the restricted HAART use and limited availability of laboratory markers in Ukraine ; the following variables were considered in the logistic regression model: HCV seropositivity, HBsAg positivity, baseline CD4 count (categorized as <200, 200–499 and ≥500 cells/μL) and timing of maternal HIV diagnosis (categorized as pre-pregnancy, first trimester, second trimester and third trimester/at delivery).
The logistic regression model investigating the likelihood of a woman having detectable HIV RNA in the third trimester/at delivery was limited to women from Western European centres and included the following variables: HCV seropositivity, HBsAg positivity, baseline CD4 cell count and antenatal ART use with incorporated timing of initiation (categorized as none, mono/dual therapy started pre-pregnancy/first trimester (‘long’), mono/dual therapy started second/third trimester (‘short’), ‘long’ HAART and ‘short’ HAART).
Standard linear mixed effects (LME) regression models were used to examine the effect of HCV coinfection on changes in log10-transformed HIV RNA levels over pregnancy . Midpoints were imputed for measurements recorded at the assay lower limit of quantification. The following variables were considered in the model in a forward stepwise selection procedure and retained if inclusion resulted in an improved log-likelihood (P≤0.10): baseline CD4 cell count, antenatal ART use with incorporated timing and timing of viral load measurement (first, second and third trimester/delivery). The final model included a random effect term for the intercept, but not for slope, as inclusion of the latter did not improve the model fit.
Statistical analyses were performed with sas software (v9.01; SAS Institute, Cary, NC, USA).
Of the 1050 HIV-infected pregnant women, 520 (49.5%) were enrolled in Ukraine and 530 (50.5%) in Western Europe. Maternal characteristics, stratified by HBsAg and HCV status, are presented in Table 1. Most women were white, although nearly a fifth (n=184) were black Africans now living in Europe, and 32% (308 of 950) were pregnant for the first time.
Table 1. Maternal characteristics by hepatitis B virus surface antigen (HBsAg) positivity and hepatitis C virus (HCV) serostatus
|Area of residence [n (%)]|
| Western Europe||23 (77)||65 (60)||15 (71)||427 (48)|
| Ukraine||7 (23)||43 (40)||6 (29)||464 (52)|
|Area of birth [n (%)]|
| Western/Central Europe||6 (20)||50 (46)||14 (67)||158 (18)|
| Eastern Europe||7 (23)||47 (44)||6 (29)||465 (52)|
| Africa||15 (50)||7 (6)||1 (5)||161 (18)|
| Other||2 (7)||1 (1)||0||19 (2)|
| Unknown||–||3 (3)||–||88 (10)|
|Age at delivery (years) [median (range)]||27.2 (20–40)||31.2 (15–43)||32.2 (25–40)||27.1 (15–43)|
|IDU [n (%)]|
| Ever||1 (3)||45 (42)||14 (67)||143 (16)|
| Never||28 (93)||57 (53)||6 (29)||725 (81)|
| Unknown||1 (3)||6 (5)||1 (4)||23 (2)|
|IDU sex partner [n (%)]|
| Ever||3 (10)||31 (29)||3 (14)||274 (31)|
| Never||25 (83)||69 (64)||17 (81)||588 (66)|
| Unknown||2 (7)||8 (7)||1 (5)||29 (3)|
|Timing of HIV diagnosis [n (%)]|
| Before pregnancy||5 (17)||56 (52)||17 (81)||396 (44)|
| During pregnancy||24 (80)||49 (45)||4 (19)||454 (51)|
| At delivery||0||2 (2)||0||27 (3)|
| Unknown||1 (3)||1 (1)||0||15 (2)|
Prevalence of and risk factors for HCV coinfection (anti-HCV antibody)
Overall, 12.3% (129 of 1050; 95% CI 10.4–14.4%) of women were HCV-seropositive, including 21 concurrent HBsAg-positive women. Prevalences of HCV seropositivity ranged from 28.9% among women who had ever injected drugs (95% CI 22.9–35.8%; 59 of 203) to 7.7% (95% CI 5.9–9.8%; 63 of 816) among women with no IDU history, and from 28.1% (95% CI 22.3–34.4%; 64 of 228) among women born in Western/Central Europe to 10.1% (95% CI 7.6–13.0%; 53 of 525) among those born in Eastern Europe and 4.4% (95% CI 1.90–8.40; 8 of 184) among those born in Africa. Logistic regression analysis identified IDU history, maternal age and area of birth as independent risk factors for HCV seropositivity (Table 2).
Table 2. Risk factors associated with hepatitis C virus (HCV) seropositivity (n=925)
|Maternal age (years)|
| <25||1.00||1.00|| |
| 25–29||2.14 (1.16–3.95)||1.67 (0.88–3.15)||0.12|
| 30–34||3.45 (1.91–6.22)||2.75 (1.43–5.29)||0.002|
| ≥35||5.26 (2.83–9.78)||3.45 (1.66–7.20)||0.001|
|Area of birth|
| Western/Central Europe||1.00||1.00|| |
| Eastern Europe||0.32 (0.21–0.49)||0.64 (0.39–1.05)||0.08|
| Africa||0.12 (0.05–0.27)||0.19 (0.08–0.45)||0.0002|
| Other||0.16 (0.02–1.20)||0.21 (0.03–1.68)||0.14|
|Injecting drug use|
| Never||1.00||1.00|| |
| Ever||4.75 (3.16–7.14)||2.92 (1.86–4.58)||<0.0001|
| No||1.00||1.00|| |
| Yes||5.61 (3.01–10.5)||5.80 (2.78–12.1)||<0.0001|
Prevalence of and risk factors for HBV coinfection (HBsAg positivity)
Overall, HBsAg prevalence was 4.9% (95% CI 3.6–6.3%; 51 of 1050), but ranged from 8.7% (95% CI 5.1–13.7%; 16 of 184) for African-born women and 8.8% (95% CI 5.4–13.2%; 20 of 228) for women born in Western or Central Europe to 2.5% (95% CI 1.32–4.20%; 13 of 525) for Eastern European women. Although maternal age was identified univariably as a risk factor for HBV coinfection, in adjusted analysis only women aged 25–29 years were at significantly increased risk (Table 3). IDU history was only a significant risk factor univariably, while having a previous IDU sex partner was negatively associated with being HBsAg positive.
Table 3. Risk factors associated with hepatitis B virus surface antigen (HBsAg) carriage (n=913)
|Maternal age (years)|
| <25||1.00||1.00|| |
| 25–29||3.75 (1.45–9.65)||2.90 (1.10–7.67)||0.03|
| 30–34||2.99 (1.11–8.10)||1.55 (0.54–4.43)||0.42|
| ≥35||3.78 (1.32–10.9)||1.70 (0.55–5.21)||0.35|
|Region of birth|
| Europe||1.00||1.00|| |
| Africa||1.91 (0.98–3.75)||2.74 (1.20–6.26)||0.02|
| Other||1.29 (0.17–10.0)||2.19 (0.25–19.1)||0.48|
| Never||1.00||1.00|| |
| Ever||1.93 (1.00–3.39)||1.70 (0.75–3.86)||0.21|
|Sex partner of IDU|
| Never||1.00||1.00|| |
| Ever||0.30 (0.13–0.72)||0.38 (0.15–0.96)||0.04|
|Hepatitis C seropositive|
| No||1.00||1.00|| |
| Yes||6.12 (3.25–11.5)||6.44 (3.08–13.5)||<0.0001|
Of the 520 women from Ukraine, 16 (3%) received no ART, 88 (17%) single-dose (sd) nevirapine (NVP) only, 393 (71%) antenatal monotherapy with or without sdNVP and 23 (4%) HAART; the proportion of women on these regimens varied significantly by HCV status, with 14% (seven of 49) of HCV-seropositive women vs. 3% (16 of 464) of HIV-monoinfected women receiving HAART, 67% (33 of 49) vs. 76% (360 of 471) on monotherapy and 18% (nine of 49) vs. 20% (95 of 464) receiving sdNVP only or no prophylaxis, respectively (χ2=12.4, P=0.002). There were no treatment differences by HBsAg status among women from Ukraine (P=0.12).
Of the 530 women from Western Europe, 51 (10%) received no antenatal ART, 159 (30%) mono/dual therapy and 320 (60%) HAART. Among this group, 43% (10 of 23), 43% (28 of 65), 47% (seven of 15) and 64% (275 of 427) of HIV/HBV-coinfected, HIV/HCV-coinfected, HIV/HCV/HBV-coinfected and HIV-monoinfected women, respectively, received HAART. In univariable logistic regression, HCV seropositivity and HBsAg carriage were associated with a significantly reduced likelihood of receipt of HAART (OR 0.42, 95% CI 0.24–0.72, P=0.002 and OR 0.34, 95% CI 0.14–0.82, P=0.02, respectively). A multivariable analysis (n=498) adjusting for baseline CD4 cell count and timing of maternal HIV diagnosis was carried out; 32 women were excluded because of missing CD4 cell count data. The significantly reduced likelihood of HAART receipt was reduced further for HCV seropositivity (AOR 0.34; 95% CI 0.20–0.58; P<0.0001), but the association with HBV did not reach statistical significance (AOR 0.57; 95% CI 0.27–1.21; P=0.14). Further analyses showed HCV seropositivity to be associated with reduced likelihood of receipt of HAART among 158 women with CD4 counts ≥500 cells/μL (AOR 0.24; 95% CI 0.08–0.69; P=0.008), but not in the larger group with CD4 counts >350 cells/μL (n=219) (AOR 0.51; 95% CI 0.23–1.14; P=0.10).
Of the 35 HIV/HCV-coinfected women on HAART, 11 (31%) received PI-based regimens, 22 (63%) NVP-based regimens (19 starting during pregnancy), and the remaining two a triple nucleoside reverse transcriptase (NRTI) regimen and an efavirenz-containing regimen (initiated after the first trimester), respectively. The median baseline CD4 count among women receiving NVP was 301 cells/μL (n=16). Two HCV-seropositive women discontinued NVP during pregnancy within 2 weeks of initiation: in one case the patient switched from NVP to saquinavir with the NRTI backbone maintained, and in the other NVP was not substituted, with the woman continuing on dual therapy [zidovudine+lamivudine (3TC)]; respective CD4 counts were 320 and 220 cells/μL. Of the 12 women with active HBV coinfection receiving dual therapy or HAART, 11 (92%) received combinations containing 3TC; no HBV-coinfected woman received tenofovir (TDF).
HIV disease: immunological and virological status
Among HIV-monoinfected women, the median baseline CD4 count was 498 cells/μL (range 301–610 cells/μL) and 9% (46 of 498; 94% of 393 women with missing CD4 cell count data were from Ukraine) had counts below 200 cells/μL; respective figures were 394 cells/μL (range 261–625 cells/μL) and 14% (10 of 73) for HIV/HCV-seropositive women, 417 cells/μL (range 216–607 cells/μL) and 22% (5/23) for HIV/HBsAg-positive women and 378 cells/μL (range 246–486 cells/μL) and 7% (one of 15) for HIV-infected women with anti-HCV antibodies and HBsAg. There was significantly more severe immunosuppression among HCV-seropositive women (Fisher's exact P<0.001) and HBsAg-positive women (Fisher's exact P<0.001) than among HIV-only-infected women.
Overall, 470 women (45%) from Western European centres had 1057 antenatal HIV RNA measurements (median 2; range 1–6). The availability of at least one measurement was not associated with HCV coinfection (P=0.11). Median HIV RNA levels were higher in every trimester for HCV-seropositive women compared with HIV-monoinfected women (Table 4). Among HCV-seropositive women, 45 had at least two HIV RNA measurements: 13% (six of 45) had undetectable HIV RNA throughout pregnancy (five on HAART and one on dual therapy) and 18% (eight of 45) had detectable levels at baseline but achieved virological suppression by delivery (six on HAART, one on dual and one on monotherapy), but most (69%; 31 of 45) had detectable levels at all measurements (15 on HAART, nine on monotherapy, five on dual therapy and two not on ART).
Table 4. HIV RNA levels, by hepatitis C virus (HCV) coinfection status and time of measurement
|HIV monoinfected||3.43 [1.40–4.45] (64)||3.23 [2.00–4.08] (201)||2.00 [1.40–3.22] (611)|
|HIV-infected/HCV-seropositive||3.65 [3.09–4.99] (19)||3.97 [2.83–4.32] (41)||2.34 [1.97–3.40] (72)|
The proportions of women with undetectable HIV RNA at their last measurement in the third trimester or at delivery were 49% (149 of 303) for the HIV-monoinfected group, 29% (13 of 45) for the HIV/HCV-coinfected group, 38% (three of eight) for the HIV/HCV/HBV-coinfected group and 42% (six of 14) for the HIV/HBV-coinfected group. Univariable logistic regression indicated that HCV seropositivity, CD4 cell count and antenatal HAART were significant predictors of detectable HIV RNA in the third trimester/at delivery. In adjusted analysis (n=370) including all these variables, women with HCV antibodies were nearly twice as likely to have a detectable viral load compared with HIV-monoinfected women (AOR 1.95; 95% CI 1.00–3.78; P=0.049).
In the LME model, HIV RNA levels depended on the time of the measurement, baseline CD4 cell count, HAART use and HCV antibody status. In the adjusted model, HCV seropositivity was significantly associated with a higher HIV RNA level (+0.28 log10 copies/mL compared with HIV-monoinfected women; 95% CI 0.04–0.52; P=0.03) and by the third trimester/at delivery, HCV-seropositive women had an estimated mean level of 2.84 log10 HIV RNA copies/mL compared with 2.56 log10 copies/mL among other women.
We report a prevalence of HCV seropositivity of 12% and HBsAg positivity of 5% in our study population of HIV-infected pregnant women living in Europe. Around a fifth of women overall were past or current IDUs and had the highest prevalence of anti-HCV antibodies, at nearly one in three. A history of IDU and older age were risk factors for HCV infection, consistent with other studies [8,12,20–23]. One in 25 African-born women was seropositive for HCV and these women were significantly less likely to be coinfected than those born in Western/Central Europe. Studies of predominantly HIV-uninfected pregnant women in sub-Saharan Africa estimate antenatal HCV prevalences of 1–5% [13,24,25] and several found no evidence of significantly increased prevalence associated with HIV infection [13,14]. The prevalence of active HBV infection found in our cohort is consistent with the lower ranges reported by other HIV cohort studies in developed country settings [4,26]. Our finding that women reporting a past or current IDU sex partner were at significantly reduced risk of being HBsAg positive may be a result of unmeasured HBV immunity, either naturally acquired or the result of vaccination.
HCV-seropositive women living in Ukraine were more likely to receive HAART than other women, although overall HAART use was substantially lower than in Western Europe . Conversely, the HCV-seropositive women living in Western Europe were 66% less likely to receive antenatal HAART. Although we found a substantially and significantly lower use of antenatal HAART associated with HCV infection in women with CD4 counts >500 cells/μL, this was attenuated for more seriously immunosuppressed women. Thus, the overall finding of less antenatal HAART use among HCV-seropositive women appears driven by a substantially reduced likelihood of receipt of HAART among women without immunological indication for HAART for their own health.
In European and US studies of HIV-infected nonpregnant adults, those with HCV-coinfection were more likely to be ARV-naïve at recruitment than those with HIV infection only, less likely to start HAART subsequently and at increased risk of discontinuing ongoing regimens [10,11,27,28]. Concerns of the physician and/or patient regarding the increased potential for HAART-associated hepatotoxicity with HCV coinfection [29,30] may partly explain the above and our study findings. Furthermore, in our study population, pregnancy may have compounded such concerns: a pregnancy-associated increased risk of hepatotoxicity, over and above that associated with female sex , was suggested by reported deaths from fulminant hepatitis among pregnant women receiving NVP-containing regimens [31–33]. Although recent studies have reported that 5–6% of pregnant women starting NVP-based HAART developed moderate to severe hepatotoxicity [34,35], one reported a 15-fold increased risk of developing NVP toxicity associated with HCV coinfection among pregnant women . However, the evidence for hepatotoxicity in HIV/HCV-coinfected pregnant women remains limited, resulting in a lack of consensus on the optimum therapeutic approach, as reflected in our findings. This appears to be particularly the case regarding the management of women with high CD4 cell counts, who are at increased risk of HAART-related toxicity and for whom HAART may not have been considered if they were not pregnant [31,37]. It was surprising that nearly a fifth of HCV-seropositive women in our study were receiving dual therapy, given current guidelines [16,37], although reassuringly these cases were concentrated in the earlier years of the study (data not shown).
Few specific recommendations exist for the management of HIV/HBV-coinfected pregnant women . Guidelines in the USA for use of ARV in HIV-infected adults now recommend combination ARV treatment for HIV/HBV-coinfected individuals when HBV treatment is indicated, regardless of CD4 cell count ; such treatment should be a combination of at least two drugs active against HBV (which include 3TC, TDF and emtricitabine) to reduce the risk of development of HBV-resistant mutations [7,39,40]. Nearly all of the treated HIV/HBV-coinfected women in our study received 3TC-containing regimens but no other anti-HBV active drug; however, this is not recommended because of the risk of selection of 3TC-resistant virus, which could be transmitted to the infant. Current guidelines recommend that HIV-infected women with chronic HBV infection should receive a three-drug regimen including a dual NRTI backbone of TDF and 3TC or emtricitabine .
We found that HCV-seropositive women (and those with HBsAg) were significantly more immunosuppressed than HIV-monoinfected women, with baseline CD4 counts on average 100 cells/μL lower in women with anti-HCV antibodies. We did not have information on time of infection nor a complete ART history for every woman, so we could not explore whether this finding was reflective of longer durations of infection and/or less effective therapeutic management. HCV genotype has been shown in some studies to be associated with HIV disease progression [41–43]. We did not have information on HCV genotype available, but it is likely that a large proportion of these HIV-infected pregnant women living in Europe had genotype 1 or 3 [38,44,45].
We investigated the impact of HCV seropositivity on viral load, particularly whether undetectable levels were reached by delivery, a key goal for prevention of mother-to-child transmission (PMTCT) [18,37], and found that this doubled the risk of having a detectable HIV RNA measurement close to delivery. HCV-seropositive women had slightly but significantly higher HIV RNA levels than other women, after adjusting for baseline CD4 cell count, ART and timing of the measurement. In an analysis from the Women and Infants Transmission Study (WITS) of 652 women enrolled before HAART became available, with an HCV prevalence of 29%, HCV coinfection was not associated with HIV RNA patterns during an average of 3 years of postnatal follow-up ; however, HCV-coinfected women were less immunosuppressed than HIV-monoinfected women in the WITS, in contrast to our results. Our finding of a significant association between HCV coinfection and uncontrolled viraemia may partly be explained by nonobserved differences between HCV-seropositive and HIV-monoinfected women, for example relating to adherence, duration of HIV infection and drug resistance.
A limitation of our study was that HBV and HCV serology was based on extraction of antenatal test results rather than a standardized serological survey; however, the study centres all had universal antenatal screening policies for HBV and HCV. A further limitation was that we were only able to look at HBsAg positivity and had no information on other serum markers for HBV, similar to other observational studies [4,26]. We were therefore unable to investigate serological evidence of past infection or prior immunization. This may explain why we found no association between IDU history and HBsAg positivity. We only had HCV antibody status information available, similar to other epidemiological studies , and were therefore unable to differentiate women with viral clearance from those with chronic HCV infection. It is estimated that around 10–25% of people with HCV infection will resolve the infection [46–48], although there is a lack of data on HCV viral clearance in HIV-infected individuals; some studies have reported lower likelihoods of HCV clearance among HIV-infected individuals, particularly those severely immunosuppressed [48–50]. Half our population were from centres in Ukraine and had no HIV RNA levels available; thus our finding of detectable viral load associated with HCV seropositivity is driven by the women from the Western European centres.
One in 20 HIV-infected pregnant women in our study were HBsAg positive. As HIV/HBV coinfection is associated with a higher prevalence of persisting HBV DNA and HBV envelope antigen (HBeAg) [40,51] which increase risk of HBV MTCT , it is essential that coinfected women and their infants are appropriately managed to prevent transmission. Pregnant HIV-infected women with anti-HCV antibodies living in Western European settings were less likely to receive HAART than HIV-monoinfected women. The former had higher HIV RNA levels throughout pregnancy and were more likely to deliver with detectable HIV than HIV-monoinfected women, independent of ART use, theoretically placing them at greater risk of transmitting HIV and HCV to their infants and of development of drug resistance. Further studies are needed to confirm these findings in HIV-infected women with documented HCV persistence in pregnancy. Research to guide evidence-based management of HIV/HCV-coinfected women is urgently needed.
We thank Drs Lucy Pembrey and Kirsty England (Institute of Child Health) for comments on the manuscript and Dr Deven Patel (Institute of Child Health) for statistical assistance with the LME models. The ECS collaborators at the participating centres were Prof. J. Levy, Dr M. Hainaut, Dr T. Goetghebuer, Dr P. Barlow and Dr Y. Manigart (Brussels); Dr V. Savasi, Dr S. Fiore, Prof. E. Ferrazzi, Dr A. Viganò, Dr V. Giacomet and Dr M. Crivelli (Milan); Prof. P. Martinelli, Dr A. Agangi, Drssa W. Buffolano, Dr R. Tiseo and Drssa M. Sansone (Naples); Dr C. Tibaldi, Dr S. Marini, Dr G. Masuelli and Dr F. Albano (Turin); Prof. I. Grosch Wörner, Dr C. Feiterna-Sperling and Dr S. Casteleyn (Berlin); Dr A. B. Bohlin, Dr S. Lindgren, Dr K. Elfgren, Dr B. Anzén and Dr K. Lidman (Huddinge and Solna); Prof. A. Mûr, Dr A. Payà, Dr M. A. López-Vilchez and Dr R. Carreras (Barcelona); Dr J. Jimenez (Madrid); Dr O. Coll, Dr S. Hernández and Dr J. Pascual (Barcelona); Dr S. Alberico, Dr M. Rabusin and Dr M. Bernardon (Trieste); Dr R. Malyuta, Dr I. Semenenko, Dr I. Shevchenko, Ms T. Pilipenko, Dr D. Richko and Ms Y. Khomout (Perinatal Prevention of AIDS Initiative, Odessa); Dr S. Posokhova, Dr T. Kaleeva, Dr A. Shelyag and Dr S. Servetsky (Odessa); Dr A. Stelmah, Dr G. Kiseleva and Dr O. A. Zalata (Crimean Republic). The ECS is a coordination action of the European Commission (PENTA/ECS 018865). The Medical Research Council (UK) Sexual Health and HIV Research Strategy Committee provided support to the ECS coordinating centre. This work was undertaken at GOSH/UCL Institute of Child Health which received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme. The Centre for Paediatric Epidemiology and Biostatistics also benefits from funding support from the MRC in its capacity as the MRC Centre of Epidemiology for Child Health. The views expressed are those of the authors and not necessarily those of the MRC or the Health Departments.