| |
Chronic Hepatitis C in Latinos: Natural History, Treatment Eligibility, Acceptance, and Outcomes
|
| |
| |
The American Journal of Gastroenterology
Volume 100 Issue 10 Page 2186 - October 2005
Ramsey C. Cheung, M.D.1, Sue Currie, M.A.2, Hui Shen, M.S.2, Samuel B. Ho, M.D.3, Edmund J. Bini, M.D.4, Bhupinderjit S. Anand, M.D.5, Norbert Bräu, M.D.6, Teresa L. Wright, M.D.2, for the VA HCV-001 Study Group*
1VA Palo Alto Healthcare System and Stanford University School of Medicine, Palo Alto, California; and VA Medical Centers, 2San Francisco, California, 3Minneapolis, Minnesota, 4New York, New York, 5Houston, Texas, 6Bronx, New York
ABSTRACT
OBJECTIVES: The natural history of chronic hepatitis C and treatment response are different between blacks and Caucasians, but little comparable data is available about Latinos.
METHODS: A cross-sectional secondary analysis to investigate differences between 421 anti-HCV-positive, treatment-naïve, HCV-viremic Latinos and 2,510 Caucasians in 24 VA medical centers enrolled in a prospective study.
RESULTS: Latinos were infected at a younger age and were less likely to have blood contact during combat, surgery, and needle stick injury, but were more frequently HIV coinfected (20.4%vs 3.9%, p< 0.0001) and prior HAV infection (39.9%vs 26.4%, p= 0.0001). Latinos were more likely to be treatment candidates, but less likely to actually initiate treatment. Liver histology (123 Latinos, 743 Caucasians) showed no difference in fibrosis or fibrosis rate, but steatosis (54.7%vs 43.2%, p= 0.038) was more common in Latinos. Eighty-eight Latinos and 481 Caucasians were subsequently treated with interferon-ribavirin: body mass index (BMI), duration of infection, baseline tests, liver histology and genotype distribution were similar. Compared with Caucasians, Latinos discontinued treatment prematurely more often (39.8%vs 28.9%, p= 0.043) and tended to have lower sustained virological response (SVR) rates (14.8%vs 22.5%, p= 0.10). Multivariate analysis found Latino race and history of recent alcohol use to be associated with early treatment discontinuation, whereas genotype and viral load but not ethnicity to be associated with SVR.
CONCLUSIONS: Latinos were infected younger, more frequently HIV coinfected, more likely to meet criteria for antiviral therapy yet less likely to initiate treatment and had a trend toward lower SVR rates than Caucasians, but not in severity of liver disease. Latino ethnicity was associated with early discontinuation but not as an independent predictor of SVR.
AUTHOR DISCUSSION
It is well recognized that African Americans and Caucasians have important differences in the natural history (3-5) and response to antiviral treatment (8-11). However, very little is known about ethnic differences between Latinos and Caucasians. The present study was carried out to examine ethnic differences between Latinos and Caucasians with chronic hepatitis C with respect to demographics, risk factors for infection, natural history of the disease, treatment eligibility, acceptance, and outcome. To the best of our knowledge, this is the first such study that prospectively assessed ethnic differences in a large cohort of almost 3,000 patients with HCV infection. Population-based surveys such as NHANES III suggests that the prevalence of HCV infection is higher among Latinos compared with Caucasians (2). Moreover, Latinos with HIV infection were more likely to be coinfected with HCV (19.3% in Latinos vs 10.9% in whites) (18). In our study, Latinos with chronic hepatitis C were more likely to be coinfected with HIV compared with Caucasians (20.4%vs 3.9%). This finding is somewhat unexpected since high-risk behaviors in Latino were either similar (e.g., IDU, greater than 50 lifetime sexual partners, homosexuality) or lower (e.g., sex with IDU) compared with Caucasians. Latinos were also more likely to be anti-HAV-positive compared with Caucasians (39.9%vs 26.4%), although prevalence of hepatitis B virus infection did not differ between the two groups (HBsAg positive 4.5%vs 5.5%).
In contrast to the present study, a retrospective chart review of chronic hepatitis C patients in the Los Angeles area found that Latinos (N = 95) were less likely to have a history of IDU compared with Whites (47%vs 54%, respectively). Among those who had a liver biopsy, Latinos (N = 66) had more rapid estimated rate of fibrosis progression than Whites (0.215/yr vs 0.084/yr, respectively), largely because Latinos had the shortest estimated duration of infection among all ethnic groups. Baseline serum liver tests (ALT, albumin, bilirubin) and percentage of patients with normal ALT level were similar between the two groups (7). However, an expanded retrospective study from the same group compared 323 Caucasians and 621 Latinos and found Latinos to have significantly higher levels of serum ALT, AST, and bilirubin and lower albumin levels. The authors again concluded that Latino patients had a faster progression of fibrosis compared to blacks or whites (6). Contrary to these findings, we did not observe in our large cohort any differences in the baseline liver tests, severity of disease assessed by liver histology, and duration of infection. In addition, we found that both groups were equally heterogeneous in the variability of their fibrosis progression rate, and there was no statistical difference between their mean fibrosis progression rates (SD) which was 0.08 stage/yr (0.06) in Latinos and 0.09 stage/yr (0.07) in Caucasian. The rate in Caucasians is similar to that with the study of Bonacini et al (7). However, our findings have to be interpreted with caution since the liver biopsies were interpreted by local pathologists, subjected to interobserver and tissue sample size variability.
Latinos were more likely to become infected at a younger age, but the difference was not clinically significant. In the present study, we found no difference between Latinos and Caucasians with regard to risk factors such as IDU and blood transfusion prior to 1990. Latinos were less likely than Caucasians to be exposed to other potential sources of infection like blood contact during combat (18.5%vs 27.4%), history of surgery (77.1%vs 83.2%), or needle stick injury (15.7%vs 22.3%). These risk factors have not been studied previously among Latino patients.
Only a small percentage of infected individuals are treatment candidates outside of large clinical trials (19). In a study of 293 HCV-infected patients seen at a metropolitan liver clinic in Cleveland, Ohio (20), as well as in studies from Veterans Administration medical centers (21, 22), the treatment eligibility rate was only about 25%. The most common reasons for not treating patients were nonadherence to pretreatment medical visits, comorbid medical or psychiatric disease, and ongoing alcohol or substance abuse. In our study, only patients who came to the clinic and consented to the study were included and hence only a small percentage of our patients were considered as nontreatment candidates on the basis of compliance. Similar to other studies, medical and psychiatric comorbidities as well as active or recent substance use were found to be the most common reasons for not initiating therapy. We found Latinos were more likely to be eligible for treatment both by the study's standardized treatment criteria and the treating physician's assessments, but only 20% were treated, a percentage similar to Caucasians. Further investigation of this phenomenon shows that there appears to be differences between Latinos and Caucasians in the behaviors and characteristics of those who initiated treatment compared with those who did not. When compared with those who did not initiate therapy, Caucasians who initiate therapy were younger, had a higher level of education and household income, and were less likely to have a history of IDU or alcohol use (remote and recent). The same trend was observed in Latinos who initiated therapy compared with those did not, but the differences were not statistically different. Factors such as alcohol use, education level, and household income may be associated with the ability to adhere to treatment and potentially explain why Caucasians that were candidates for treatment were more likely to initiate therapy than Latinos who were candidates.
In a recent study on a total of 45 Hispanic subjects in two treatment arms of consensus interferon monotherapy (9 mug or 15 mug TIW), the SVR in Hispanics (12%) was intermediate between whites (24%) and African Americans (4%). However, in patients with genotype 1, the SVR was 15% in whites and 13% in Hispanics (p= NS). In genotype 2, the SVR was 41% in Whites but only 10% in the 10 Hispanics (23). By pooling data from two different investigator-initiated trials using nonstandard doses of combination therapy (interferon and ribavirin), Hepburn et al. found the overall SVR in Latinos was intermediate to that of Whites and Blacks (12). When ethnicity was used as predictor of response for Latinos as compared with Whites in logistic regression analysis, the odds ratio (OR) (95% confidence interval (CI)) was 0.5 (0.3-1.0). In the multiple logistic regression analysis, independent predictors of favorable response were sex (female), genotype (2 or 3), liver histology (mild), and no dose reduction during treatment.
Latinos were more likely to discontinue treatment early despite similar baseline WBC counts and hemoglobin. We found Latinos had a lower EOT and SVR rate to standard interferon and ribavirin therapy compared with Caucasians, but the difference did not reach statistical significance, and Latino ethnicity was not an independent predictor of response in our multivariate regression analysis. The SVR was much lower in Latinos than Caucasians in nongenotype 1 (p= 0.056) but was similar in genotype 1. There were several potential explanations for the reduced response rate among Latinos. Latinos were more likely to be genotype 1, higher viral load, and higher rate of early discontinuation of treatment. Among Latino who achieved EOT response, almost half relapsed as compared to only one-third in Caucasian. Finally, Latinos were more likely to have steatosis on liver biopsy, which has been found to be associated with reduced SVR in some studies but not in the current analysis. Multivariate analysis identified only genotype 1 and high viral load as associated with reduced SVR rate and eliminated race as a potential factor in this respect. However, it is possible that there could be differences between various ethnic subgroups due to the diverse racial background among Latinos.
In summary, we found that major demographics, risk factors such as transfusion and IDU, duration of infection, baseline serum liver function tests, genotype distribution, severity of liver disease, and fibrosis progression rate were similar between the two groups. However, Latinos had a lower prevalence of potential risk factors such as other types of blood exposure, but were more likely to be coinfected with HIV infection. Interestingly, Latinos were more likely to be treatment candidates yet less likely to initiate treatment. There was a trend toward lower sustained virological response to antiviral therapy in Latinos but differences did not reach statistical significance. In multivariate analysis, Latino ethnicity was an independent predictor of earlier discontinuation, but not response. Further study should be performed to determine barriers to initiate antiviral therapy and reasons for early treatment discontinuation in Latinos.
RESULTS
Study Population
A total of 4,462 patients were enrolled in the study and 813 (18.6%) were started on treatment. In the screening phase, there were 2,510 Caucasians and 421 Latinos, of which 481 (19.2%) Caucasians and 88 (20.9%) Latinos were treated with interferon and ribavirin. These patients form the basis of the present study. There were no differences between Latinos and Caucasians with respect to age, alcohol consumption pattern, other socioeconomic factors, and service record. Previous exposure to hepatitis A virus and HIV coinfection were more common in Latinos. The estimated mean duration of infection and standard deviation (SD) was 26.9 (8.8) yr in Latinos and 26.1 (8.4) in Caucasians (p= 0.065). However, Latinos were infected at a younger age than Caucasians (22.9 (8.8) vs 24.2 (10.1), p= 0.008). The most common risk factor for HCV infection in both groups was injection drug use (IDU), seen more frequently in Latinos than Caucasians (p= 0.058). However, sexual contact with injection drug user and other forms of blood contact such as during combat, needle stick injury, and surgery were all less common in Latinos (p< 0.05).
Treatment Eligibility and Decision
Latinos were more likely to meet all treatment criteria (48.9%vs 39.6%) and to be considered a treatment candidate by their treating physician (43%vs 37.8%). Despite the higher percentage of eligible Latino candidates compared to Caucasians, we found no difference in the percentage of Latino patients who actually initiated the treatment (20.9% and 19.2%).
The main contraindications at the time of initial screening for antiviral therapy were similar in both groups. Recent or ongoing substance use and psychiatric comorbid conditions were the most frequent reasons considered by the physicians for not initiating treatment. Another common exclusionary criterion was persistently normal ALT, which was seen more frequently in Caucasians compared with Latinos. When the untreated and treated groups were compared, Caucasians who did not initiate treatment were significantly older (age of HCV infection and age at time of treatment evaluation), had lower level of education and household income, and were more likely to have history of IDU and remote or recent alcohol use. A similar trend was observed in the Latino group who did not begin treatment compared with those who initiated therapy, but unlike in the Caucasian group, the differences were not statistically different.
Liver Histology
Thirty percent of the enrolled patients in each group underwent liver biopsy. Cirrhosis was less common in Latinos (11%) than Caucasians (15.6%). The degree of inflammation was similar in both groups, but steatosis was seen more frequently in Latinos (p= 0.038). The mean (SD) fibrosis progression rate, defined as the fibrosis stage on liver biopsy over estimated duration of infection, was 0.08 stage/yr (0.06) in Latinos and 0.09 stage/yr (0.07) in Caucasian, p= 0.14.
Characteristics of Treated Patients
A total of 88 Latinos and 481 Caucasians were treated according to standard guidelines. There were no differences between the two treatment groups in terms of duration of infection, baseline weight, body mass index (BMI), serum ALT, hemoglobin, white blood cell (WBC), platelet count, alpha feto-protein (AFP), ferritin, and frequency of genotype 1 infection (74.2% in Latinos vs 68.7% in Caucasians, p= 0.36). The mean (SD) ALT was 100.4 (56.2) and 101.5 (56) in Latinos and Caucasians, respectively. In Latinos, the baseline absolute neutrophil count was lower (p= 0.015) and the viral load was higher (p= 0.03) compared with Caucasians. Only 3.5% of the Latinos who received treatment had cirrhosis on liver biopsy (defined as stage 4 fibrosis) as compared to 16.8% in Caucasians (p= 0.009), but the percentage of patients with advanced fibrosis (stage 3 or 4), and the mean grade and stage on pretreatment liver biopsy were not statistically different between the two groups.
Treatment Outcomes
Although not statistically significant, we found weak evidence for an association between lower EOT and SVR rates in Latinos. The EOT response (27.3%vs 37.0%, p= 0.08) and SVR rates (14%vs 22.5%, p= 0.11) were lower in the Latinos compared with Caucasians, respectively.
Among genotype 1 patients, the SVR was 10.2% in Latino and 14.6% in Caucasians (p= 0.14). In nongenotype 1, the SVR was 17.7% in Latino and 38.4% in Caucasian (p= 0.055). A higher proportion of Caucasians (57.8%) than Latinos (44.3%) were able to complete the treatment without any dose reduction. Latinos were more likely to discontinue treatment early either due to sideeffects or dropout: 39.7%vs 28.9%(p= 0.043). In univariate analysis, in addition to race, other factors associated with early discontinuation were weight <75 kg (p= 0.009), BMI <28.6 (median) (p= 0.039), baseline hemoglobin <13 g/dL (p= 0.011), average alcohol use of three or more drinks per day in the past 12 months (p= 0.0001), and household income of less than $25,000 (p= 0.019). In the multivariate analysis, only average alcohol use of 3 or more drinks per day in the past 12 months (p= 0.0001) and Latino race (p= 0.040) were associated with early discontinuation of treatment.
Factors Predictive of Treatment Outcome
Univariate analysis found the following factors to be predictors of response: genotype (1 vs 2 or 3, OR = 0.29; 95% CI = 0.18-0.46), viral load (high vs low, OR = 0.56, 95% CI = 0.36-0.88), and AFP (per ng/mL, OR = 0.94, 95% CI = 0.89-0.99), but not race (Latino vs Caucasian, OR = 0.60, 95% CI = 0.32-1.12). Multivariate analysis only found genotype 1 (OR 0.196; 95% CI = 0.099-0.390) and high viral load (OR 0.439; 95% CI = 0.221-0.872), but not Latino ethnicity, as factors associated with lower SVR.
METHODS
Study Design
This was a multicenter two-phase prospective study. The first (screening) phase of the study evaluated the epidemiology of HCV infection and treatment candidacy among United State veterans. The second (treatment) phase included all subjects who were treated with interferon alpha-2b and ribavirin. Patients were recruited from the specialty clinics of 24 geographically diverse VA medical centers throughout the United States between December 1999 and December 2000. All patients had a positive test for anti-HCV (Ortho HCV ELISA version 2.0 or 3.0; Ortho-Clinical Diagnostics, Inc., Raritan, NJ) and were considered for treatment with interferon alfa-2b and ribavirin. All subjects were consented and were administered an extensive questionnaire by trained study staff. Additional information such as lab values was obtained from patient medical records. Analyses were performed according to the completed questionnaires and available data, and only patients who were viremic as documented by a positive HCV RNA by polymerase chain reaction (PCR) were included in the study. The estimated time of HCV infection was determined by the first exposure to a risk factor (e.g., injection drug use, transfusion). Liver biopsies were performed on selected patents at the discretion of the treating physicians and were interpreted by the local pathologists using standard classification. Race and ethnicity was determined by self-reporting as African American, Asian, Caucasian, Latino, Native American, or other. In this study, the terms Hispanic, Mexican American, and Latino were used interchangeably, as well as African American and black, and non-Hispanic white and Caucasian. Further details of this study will be described elsewhere. The study was approved by the local institutional review boards of each medical center.
Treatment Eligibility
The treatment candidacy of all enrolled patients was determined following recommendations of the VA's HCV Treatment Guidelines at the time of the study (2/25/99 version) (16), which were consistent with the 1997 National Institutes of Health (NIH) consensus conference (17). Patients with chronic hepatitis C were eligible for treatment if they were hepatitis C treatment naïve, had compensated liver disease, had normal values of serum creatinine, and met predefined laboratory parameters (total bilirubin <2.0 g/dL, hemoglobin >=12 g/dL, neutrophils >=1,500/mm3, platelets >=85,000/mm3). In addition, the study allowed the treating clinician to determine whether the patient was considered a treatment candidate independent of the VA Treatment guidelines.
Treatment
The standard therapy at the time of the study was interferon alfa-2b (IFN) 3 million units (MU) three times per week (TIW) plus ribavirin (RBV) 1,000-1,200 mg/day (cutoff at 75 kg weight) (Rebetrontm, Schering-Plough, NJ). The duration of treatment was 48 wks for genotype (GT)-1 or -4 and 24 wk for GT-2 or GT-3. Therapy was discontinued at wk 24 in patients with GT-1 or -4 who remained viremic. All patients were followed during antiviral therapy and for another 24 wks after completion of therapy. As per the recommended guidelines, the dose of IFN was reduced from 3.0 to 1.5 MU TIW for a neutrophil count of <750/mm3 or a platelet count of <50,000/mm3, and the treatment was discontinued for a neutrophil count of <500/mm3 or a platelet count of <20,000/mm3. Similarly, the RBV dose was reduced from 1,000-1,200 mg/day to 600 mg/d for a hemoglobin level of <10.0 g/dl, and RBV was stopped with a hemoglobin concentration of <8.5 g/dl. SVR was defined as undetectable plasma HCV RNA (<100 copies/mL) by qualitative PCR assay (Amplicor®, Roche Diagnostics, Nutley, NJ) 24 wk after the end of treatment. Treatment outcome was analyzed on an intention-to-treat basis.
Statistical Analysis
Statistical analysis was performed using SAS software (SAS Institute, Cary, NC). Continuous variables were compared using the Student's t-test or the Mann-Whitney U-test as appropriate. Categorical variables were compared by chi2analysis or Fisher's exact test where appropriate. Bivariate analysis was conducted when we were analyzing one predictor (race) and one outcome (treatment). Logistic regression analyses were performed to determine association of baseline variables with early discontinuation and SVR. Subsequently, multivariate logistic regression models were created involving all statistically significant variables identified in the univariate analyses. The relatively restrictive inclusion criterion for our study further reduced the risk of type 1 error. A 2-tailed p value of <0.05 was considered statistically significant. For the multivariate logistic regression analyses, interaction effects were assessed to ascertain whether certain independent variables interacted and affected the outcomes differently than they did independently. It was determined that there was a significant interaction effect between the variables Latino and income. New variables, Latino with income <$25,000 and Latino with income >=$25,000, were created to address this effect.
|
|
| |
| |
|
|
|
