Factors associated with guideline-based hepatitis C virus (HCV) treatment initiation in HIV/HCV-coinfected patients: role of comorbidities and physicians' perceptions
Correspondence: Dr Firouzé Bani-Sadr, Department of Infectious Diseases, Robert Debré hospital, University of Champagne Ardenne, 51 092 Reims, France. Tel: 3 26 78 78 94; fax: 3 26 78 34 46; e-mail: firstname.lastname@example.org
Many HIV-infected patients with chronic hepatitis C virus (HCV) infection do not receive treatment for HCV infection, often because of contraindications or poor adherence to anti-HIV therapy. The aim of this study was to identify factors influencing guideline-based HCV treatment initiation in a large cohort of HIV/HCV-coinfected patients.
Between 2005 and 2011, 194 (40.5%) of 479 coinfected patients not previously treated for HCV infection started this treatment based on current recommendations, i.e. a Metavir score > F1 for liver fibrosis; HCV genotype 2 or 3 infection; or HCV genotype 1 or 4 infection and low HCV viral load (< 800 000 IU/mL), whatever the fibrosis score. Clinical and biological data were compared between patients who started HCV therapy during follow-up and those who did not.
In multivariate analyses, good adherence to treatment for HIV infection, as judged by the patient's physician, was associated with HCV treatment initiation [odds ratio (OR) 2.37; 95% confidence interval (CI) 1.17–4.81; P = 0.017], whereas patients with children (OR 0.53; 95% CI 0.30–0.91; P = 0.022) and those with cardiovascular disease or respiratory distress (OR 0.10; 95% CI 0.01–0.78; P = 0.03) were less likely to be treated.
Adherence to treatment for HIV infection, as judged by the patient's physician, appears to have a major influence on the decision to begin treatment for HCV infection in coinfected patients. This calls for specific therapeutic education and adherence support in order to ensure timely anti-HCV therapy in this population.
Compared with isolated hepatitis C virus (HCV) infection, concurrent infection by HIV and HCV results in more rapid progression towards cirrhosis and liver failure [1, 2]. In the last decade, chronic liver disease, mainly resulting from HCV infection, has become one of the leading causes of morbidity and mortality in HIV-infected patients . In the first published trials of anti-HCV therapy in this setting, the pegylated interferon alpha (peg-interferon) and ribavirin combination was less effective in HIV-coinfected patients. Sustained virological response (SVR) rates ranged from 14 to 38% among patients with HCV genotype 1 infection and from 44 to 73% among those with genotype 2 or 3 infection [4, 5]. However, markedly higher SVR rates among coinfected patients, reaching more than 50%, have been observed in clinical practice in the last few years . Furthermore, SVR has been found to have major benefits in HIV/HCV-coinfected patients, with slower progression of fibrosis, less frequent severe liver toxicity of antiretroviral therapy, and a reduction in liver-related complications and mortality [1, 7]. Under-treatment of chronic HCV infection in HIV-infected patients remains frequent, however, for several reasons, including contraindications, patient refusal and poor adherence [6, 8, 9]. The French Expert Group recommendations for medical management of HIV infection advised the use of peg-interferon plus ribavirin combination therapy for patients with stage ≥ F2 fibrosis and, whatever the fibrosis score, for patients with a high likelihood of achieving an SVR, i.e. patients infected with genotype 2 or 3 and those infected with genotype 1 or 4 who have a low viral load (< 800 000 IU/mL) . The aim of this study was to identify factors associated with guideline-based HCV treatment initiation in a large cohort of HIV/HCV-coinfected patients.
This study involved HIV/HCV-coinfected patients enrolled in the French National Agency for Research on AIDS and Viral Hepatitis (ANRS) CO13 HEPAVIH cohort, a French prospective hospital-based cohort of HIV/HCV-coinfected adults, created in 2005 and involving 17 hospital units, of which 16 were academic hospitals .
The outcome variable was the initiation of HCV treatment, defined as the first prescription of peg-interferon and ribavirin to an HCV-treatment-naïve patient during follow-up (between 2005 and 2011). In a first analysis, clinical and biological data were compared between patients with an indication for HCV therapy and who started treatment during follow-up (patients with ≥ F2 fibrosis, patients infected with genotype 2 or 3, and patients infected with genotype 1 or 4 who had a low viral load (< 800 000 IU/mL), whatever the fibrosis score), and patients who remained HCV-treatment-naïve. The second analysis was restricted to patients with a formal indication for HCV therapy, i.e. a fibrosis score ≥ F2 regardless of the HCV genotype.
Liver fibrosis was assessed by liver histology or noninvasive methods [transient elastometry (TE) or FibroTest (BioPredictive, Paris, France)]. The results of liver biopsy (LB), when available, were documented and graded with the Metavir system. Liver fibrosis was systematically assessed with noninvasive methods [FibroTest and/or FibroScan (Echosens, Paris, France)]. FibroScan cut-offs used for conversion to the Metavir score were as follows: F0–F1, < 7.1 kPa; F2, 7.1–9.5 kPa; F3, 9.5–12.5 kPa, and F4, ≥ 12.5 kPa . Only liver stiffness measurements based on 10 successful acquisitions with a success rate of at least > 60% and an interquartile range of < 30% were considered reliable.
The following data were collected at enrolment, during follow-up, and at HCV treatment initiation: age, sex, risk categories for HIV and HCV infection, alcohol consumption and smoking, active drug use (yes/no), duration of HIV infection, use of highly active antiretroviral therapy (HAART) and duration of antiretroviral therapy, duration of HCV infection (from the date of the first transfusion, first narcotic injection, or first positive HCV serology in subjects infected through sexual contact and who did not develop acute hepatitis), liver enzyme activities, CD4 lymphocyte counts (absolute number and percentage) at baseline and at initiation of antiretroviral therapy, plasma HIV RNA and HCV RNA load, and the HCV genotype.
The self-administered patient questionnaire used at baseline included items on socio-demographic characteristics (secondary school certificate, parenthood, living alone/with a partner, employment, and housing conditions), psychiatric disorders (depressive symptoms) and use of antidepressant drugs. Items on patients' beliefs about the effectiveness and toxicity of HCV treatment, measured on a four-point visual analogue scale, were also included .
Depressive symptoms were assessed with the Center for Epidemiological Studies Depression Questionnaire (CES-D), which provides a global depression score ranging from 0 to 60, with gender-specific cut-off values (17 for men and 23 for women) [14, 15]. Score values above these cut-offs were taken to indicate the presence of depressive symptoms.
A section of the self-administered questionnaire, based on the Symptoms Distress Module proposed by Justice et al., collected data on the occurrence of 39 treatment-related symptoms (defined here as self-reported side effects) over the previous 4 weeks, and the discomfort they caused .
A self-administered questionnaire was provided to the physician for each one of his/her patients which allowed the collection of data on the physicians' experience with HIV and the physicians' perceptions about their patients in terms of adherence to HIV treatment, depression, and alcohol use at enrolment and at scheduled annual visits.
Among patients potentially eligible for HCV treatment, we compared clinical, biological and psychosocial data between patients who initiated HCV therapy during cohort follow-up and patients who did not begin this treatment. Explanatory variables were collected at the HCV treatment initiation visit for patients who had initiated treatment, and at the last follow-up visit for HCV-treatment-naïve patients. Continuous variables were expressed as medians (range), and categorical variables as frequencies and percentages. The χ2 test or Fisher's exact test was used to compare qualitative variables, and the Mann–Whitney–Wilcoxon test was used for quantitative variables. Logistic regression models were used to identify associations between the outcome variable and explanatory variables. Factors with P-values < 0.20 in univariate analyses were included in a multivariate model. The Hosmer–Lemeshow backward procedure was used to select factors associated with HCV treatment initiation. All statistical tests were two-sided, with a type I error of 5%.
Among 1048 chronically HCV/HIV-coinfected patients included in the HEPAVIH cohort, 347 patients (33.1%) had already received HCV treatment before enrolment in the cohort (before 2005) and were excluded from this study. Compared with all remaining patients, these previously treated patients were significantly older [median age 46 (range 31–81) years vs. 44 (range 18–70) years for remaining patients; P = 0.0023], had a longer duration of antiretroviral therapy [median 125 (range 1–255) months vs. 110 (range 1–244) months, respectively; P = 0.0001], had a lower prevalence of alcohol consumption [never or former (as opposed to current) drinkers, 28% vs. 20%, respectively; P = 0.0062], had a more frequent liver fibrosis score ≥ F2 (66% vs. 47%, respectively; P < 0.0001), and had more frequent cirrhosis (35% vs. 19%, respectively; P < 0.0001).
Among the 479 patients with an indication for HCV therapy at enrolment in the cohort, 194 (40.5%) initiated this treatment during follow-up. Table 1 shows the last clinical and biological data obtained before HCV treatment initiation in patients who started HCV therapy (n = 194) and in patients who did not start HCV therapy (n = 285).
Table 1. Last clinical and biological data before guideline-based hepatitis C virus (HCV) treatment initiation in HIV/HCV-coinfected patients (indications: ≥ F2 fibrosis, genotype 2 or 3 infection, and genotype 1 or 4 infection with low viral load (< 800 000 IU/mL)
|Number (total = 479)||285||194|| |
|Age (years)||48 (21–75)||45 (23–62)||<0.0001|
|Male gender||192 (67)||145 (75)||0.08|
|European origin||176 (69)||127 (72)||0.72|
|High-school certificate (yes)||60 (28)||54 (36)||0.10|
|Living with a partner||117 (49)||85 (50)||0.91|
|Having children||88 (37)||46 (27)||0.03|
|Employment (yes)||108 (45)||87 (51)||0.27|
|Home owner or tenant (yes)||187 (79)||131 (77)||0.58|
|Alcohol consumption|| || ||0.03|
|None or formerly||107 (43)||91 (52)|| |
|≤50 g daily for men or ≤30 g daily for women||132 (53)||71 (41)|| |
|>50 g daily for men or >30 g daily for women||11 (4)||13 (7)|| |
|Duration of HIV infection (years)||21 (2–29)||18 (1–29)||<10-4|
|AIDS||92 (32)||55 (31)||0.70|
|Lipodystrophy (yes)||118 (52)||91 (53)||0.84|
|CD4 count nadir (cells/μL)||138 (1–1258)||144 (0–918)||0.12|
|Absolute CD4 count (cells/μL)||490 (6–2574)||444 (63–1363)||0.16|
|CD4 count <150 cells/μL||21 (7)||4 (2)||0.01|
|Patients with <50 HIV RNA copies/ml||211 (76)||143 (78)||0.58|
|Ongoing antiretroviral treatment*||263 (92)||179 (92)||0.18|
|Duration of antiretroviral treatment (months)*||146 (0–268)||122 (0–232)||<10-4|
|Good adherence to HIV therapy, as perceived by physician (yes)||181 (78)||130 (84)||0.04|
|Duration of HCV infection (years)||27 (2–40)||24 (1–38)||<10-4|
|HCV infection through injecting drug use||212 (84)||139 (80)||0.36|
|History of decompensated cirrhosis, hepatocarcinoma or transplantation (%)||10 (4)||7 (4)||0.95|
|Metavir score > F1||180 (68)||147 (77)||0.04|
|HCV viral load (IU/ml), log10||5.9 (2.1–7.7)||6.0 (3.1–7.8)||0.07|
|HCV genotype|| || ||0.01|
|1||128 (45)||106 (55)|| |
|2||16 (6)||10 (5)|| |
|3||66 (23)||51 (26)|| |
|4||75 (26)||27 (14)|| |
|ALT (× normal upper limit)||1.2 (0.1–27.7)||1.4 (0,3–9,9)||0.005|
|Platelet count (cells/μL)||192 (23–589)||183 (32–458)||0.59|
|Prothrombin time <70%||10 (4)||11 (6)||0.22|
|Patients with negative perceptions of HCV treatment side effects (severe vs. moderate or none)||67 (49)||47 (41)||0.47|
|Patients with negative perceptions of HCV treatment efficacy (none or low vs. high)||52 (38)||48 (42)||0.85|
|Type 2 diabetes||10 (4)||12 (6)||0.17|
|Cardiovascular disease or respiratory distress||21 (7)||8 (4)||0.14|
|History of multiple treatments for depression, as reported by physician||53 (19)||29 (15)||0.30|
|Depression: patient's self perception||152 (64)||106 (62)||0.68|
|Follow-up in hospital units having enrolled more than 25 patients||225 (79)||149 (77)||0.58|
In univariate analyses, the following factors were significantly related to HCV treatment initiation: age, male sex, less alcohol consumption (current vs. former or none), childless status, shorter durations of HIV infection and antiretroviral therapy, a higher current CD4 cell count (≥ 150 vs. < 150 cells/μL), a shorter duration of HCV infection, a higher alanine aminotransferase (ALT) level, and better adherence to HIV treatment, as perceived by the physician.
No association was observed between HCV treatment and ethnicity, living alone/with a partner, employment status, home ownership/rental, HCV infection through injecting drug use, AIDS status, ongoing antiretroviral treatment, lipodystrophy, a history of decompensated cirrhosis, hepatocarcinoma and/or transplantation, type 2 diabetes, a history of multiple treatments for depression (as reported by the physician), depressive symptoms or negative perceptions about HCV treatment efficacy or side effects and the hospital unit in which the patients were followed up.
Other factors such as a higher level of education (P = 0.098), a higher CD4 cell nadir (P = 0.099) and no history of cardiovascular disease or respiratory distress (P = 0.15) tended to be associated with HCV treatment initiation.
The liver fibrosis score, HCV genotype and HCV viral load (the parameters used to select the study population) were not included in the multivariate model.
In multivariate analysis (Table 2), good adherence to antiretroviral treatment, as perceived by the physician [odds ratio (OR) 2.37; 95% confidence interval (CI) 1.17-4.81; P = 0.017], remained associated with HCV treatment initiation, whereas parenthood (OR 0.53; 95% CI 0.30-0.91; P = 0.022) and a history of cardiovascular disease or respiratory distress (OR 0.10; 95% CI 0.01-0.78; P = 0.03) were associated with the absence of HCV treatment.
Table 2. Correlates of hepatitis C virus (HCV) treatment initiation in HIV/HCV-coinfected patients with an indication for HCV therapy* (n = 479)
|Male gender||1.43 (0.95–2.15)||0.08|| || |
|High-school certificate (yes)||1.46 (0.93–2.28)||0.098|| || |
|Having children||0.63 (0.41–0.97)||0.036||0.53 (0.30–0.91)||0.022|
|Alcohol consumption: current vs. past or never (%)||0.67 (0.46–0.99)||0.045|| || |
|CD4 count <150 vs. ≥150 cells/μL||3.73 (1.26–11.1)||0.017|| || |
|ALT (× normal upper limit)||1.08 (0.98–1.20)||0.13|| || |
|Good adherence, as perceived by the physician (yes)||1.50 (0.87–2.55)||0.14||2.37 (1.17–4.81)||0.017|
|Cardiovascular disease or respiratory distress||0.54 (0.23–1.25)||0.15||0.10 (0.01–0.78)||0.03|
As adherence to HIV treatment was associated with HCV treatment initiation, we examined whether certain patient characteristics were associated with good adherence. Patients with good adherence to HIV treatment, as perceived by the physician, had lower alcohol consumption [current (as opposed to former or none), 32% vs. 51% for patients with poor perceived adherence; P = 0.0013] and fewer depressive symptoms (60% vs. 75%, respectively; P = 0.02), were more likely to own or rent their home (80% vs. 67%, respectively; P = 0.03), were less likely to have contracted HCV through injecting drug use (81% vs. 91%, respectively; P = 0.048) and had a higher CD4 cell nadir [median 156 (range 64–266) cells/μL vs. 98 (range 30–184) cells/μL, respectively; P = 0.002] than patients with poor perceived adherence. HIV viral load did not differ between the two groups.
We then restricted the analysis to the 339 patients with a formal indication for HCV therapy (fibrosis score ≥ F2 whatever the HCV genotype), of whom 147 patients (43.3%) received this treatment. In multivariate analysis, only good perceived adherence to HIV treatment remained associated with HCV treatment initiation (OR 2.40; 95% CI 1.08-5.32; P = 0.031).
We analysed factors associated with HCV treatment initiation between 2005 and 2011 in a large cohort of HIV/HCV-coinfected patients (n = 1048) managed in 17 French hospital units. One-third of the patients (n = 347) had been treated for HCV infection before their enrolment in the cohort (before 2005). Among the remaining, treatment-naïve patients, 194 (40.5%) of those with an indication for HCV therapy, as defined in French guidelines (see Methods), started treatment with peg-interferon plus ribavirin during cohort follow-up . [European guidelines for HCV therapy are similar to French guidelines, except that they include a lower HCV viral load threshold (< 400 000–500 000 IU/mL) for patients with genotype 1 infection, and do not deal with HCV genotype 4 infection ].
Patients with good adherence to HIV treatment (as perceived by their physician), patients who were childless, and patients with no history of cardiovascular disease or respiratory distress were more likely to be treated. When the analysis was restricted to the 339 patients with a formal indication for HCV treatment (fibrosis score ≥ F2), the only factor associated with HCV treatment initiation was good perceived adherence to HIV treatment.
Good adherence is critical for the efficacy of both anti-HIV and anti-HCV therapy [17, 18]. High levels of adherence are required to obtain sustained HIV suppression and long-term immunological and clinical benefits. The optimal level of adherence may differ according to the antiretroviral class, and should be better than 95% for nonboosted or boosted protease inhibitor (PI) regiments and at least 80% for nonnucleoside reverse transcriptase inhibitor (NNRTI) regimens [18, 19]. Good adherence to HCV therapy is also required to achieve a sustained viral response, with a ‘golden rule’ of 80/80/80 (continued prescription of at least 80% of the interferon dose and 80% of the ribavirin dose for at least 80% of the planned treatment period) . Patients who take > 80% of their peg-interferon alpha-2b plus ribavirin dosage for the entire treatment period have higher SVR rates [17, 20]. Adherence tends to be better for peg-interferon than for ribavirin, but decreases over time for both medications. Self-reported adherence is usually higher than that assessed by electronic measures, and the degree of discrepancy increases as treatment progresses .
In our study, 16% of patients did not receive HCV treatment because their physician considered their adherence to HIV treatment to be poor. In other studies, poor adherence or patient refusal was the reason for noninitiation of HCV treatment in 18% to 22% of patients [8, 9]. Specific interventions should be able to reduce this figure.
Social determinants may have a bearing on an individual's decision to begin anti-HCV treatment. In our study, parenthood was the only such factor associated with a lower rate of HCV therapy. Living with a partner was not a barrier to HCV therapy, but there was a trend in univariate analysis towards a higher rate of HCV treatment initiation in men (P = 0.08). The long duration of HCV therapy and side effects such as asthenia and depressive symptoms probably explain this barrier to HCV treatment in mothers. Newly available anti-HCV drugs with better efficacy, possibly allowing shorter treatment, may help to overcome this barrier. Housing status was not directly associated with HCV treatment initiation in our study, but patients who did not either own or rent their home were more likely to be considered as having poor adherence to HIV therapy by their physicians. Social interventions may help to ensure timely HCV treatment in this population.
Barriers to HCV treatment are more frequent in HIV/HCV-coinfected patients than in patients with isolated HCV infection, and mainly involve severe psychiatric disorders and ongoing alcohol use . We found that cardiovascular disease and respiratory distress were also barriers to HCV treatment initiation. Patients with a history of multiple treatments for depression were less likely to be treated when treated and nontreated patients were compared, whatever the indication (OR 0.39; 95% CI 0.17-0.88; P = 0.02) (data not shown).
Interestingly, despite the lower efficacy and poorer tolerability of HCV treatment in HIV/HCV-coinfected patients, we found no association between patients' negative perceptions of side effects or efficacy and treatment initiation. This could be a result of the fact that most patients (68%) had a formal indication for HCV treatment (Metavir score ≥ F2).
The main strength of our study is its large size: we enrolled 1048 HIV/HCV-coinfected patients managed in 17 hospital units throughout France, all with lengthy follow-up. This population was therefore probably adequately represented in the sample. The most important limitation of this study is probably our inability to determine the main reasons why physicians decided not to start HCV therapy. The number of different reasons cited by physicians for nontreatment has fallen over the years. HCV treatment is deemed less questionable: the lack of liver biopsy no longer seems to be a major barrier, and the number of contraindications to HCV treatment has also decreased . However, the numerous side effects associated with either peg-interferon or ribavirin, such as asthenia or depression, and/or the increase in the number of pills, which may result in poor adherence to therapy and therefore poor efficacy, could remain the reasons why physicians defer HCV treatment in some patients.
The second limitation of our study is that adherence to HIV therapy was only evaluated through the physician's personal perception and not by more objective methods such as self-reporting, pharmacy refill data or pill counting. However, in our study, perceived poor adherence to HIV therapy was associated with well-known factors such as alcohol consumption, homelessness, a history of injecting drug use, and depressive symptoms [22, 23]. The nadir CD4 cell count was also lower in patients with poorer adherence, reflecting a delay in linkage to care. In contrast, the rates of undetectable plasma HIV and CD4 cell counts were similar between patients who started HCV therapy and those who did not, maybe reflecting a hint of a discrepancy between the physician's personal perception and actual antiretroviral therapy adherence. Therefore, we presume that the physician's perception about adherence may be more related to perceived tolerability and potential risk of HCV treatment interruption. Developing expertise in HCV treatment and effective strategies for managing the side effects often encountered during HCV treatment might help to change physicians' personal perception of patients' poor adherence and help them to overcome the challenges of HCV therapy.
In conclusion, adherence to HIV therapy, as perceived by the patient's physician, seems to play a key role in the decision to prescribe treatment for HCV infection to HCV/HIV-coinfected patients. As a high level of adherence is required for new HCV PIs, which carry a higher risk of inducing resistance  and are possibly more toxic, developing the physician's expertise in HCV treatment, therapeutic education and adherence support might help to enable timely HCV treatment in patients with poor adherence to HIV therapy.
This study was sponsored and funded by the French National Agency for Research on Aids and Viral Hepatitis (ANRS), with the participation of Glaxo-SmithKline, Roche and Schering-Plough.