| |
Do HIV Care Providers Appropriately Manage Hepatitis B in Coinfected Patients Treated with Antiretroviral Therapy?
|
| |
| |
Clinical Infectious Diseases April 1, 2007
Mamta K. Jain, Christopher K. Opio, Chukwuma C. Osuagwu, Rathi Pillai, Philip Keiser, and William M. Lee
Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas
"....In the United States, up to 10% of all HIV-infected persons are coinfected with hepatitis B virus (HBV)... When patients receive drugs with dual activity (e.g., lamivudine, tenofovir disoproxil fumarate, and emtricitabine), they may experience acute flares of hepatitis resulting from immune reconstitution, drug withdrawal, the development of HBV drug-resistance mutations, or all 3 causes. The increased risk of hepatocellular carcinoma (HCC) [13] and the emergence of HBV drug-resistant mutations necessitates careful monitoring of coinfected patients after antiviral therapy is begun.....the goal of our study was to determine the level of care for HBV infection that was being provided to patients with HIV infection by their health care providers.... Our data strongly indicate that, during the period from 1999 to 2003, HIV care providers frequently failed to initially evaluate or to subsequently monitor HBV infection status in coinfected patients despite careful follow-up for HIV infection. 92% of this patient group received ART regimens that contained 1 HBV-active nucleosides and/or nucleotides. Although every patient is tested for HBsAg, anti-HCV, and antibodies for hepatitis A, only 16% had had an HBV load measured to assess the status of disease and the possible need for treatment of HBV infection prior to initiation of ART. Assessing baseline viral and serologic markers of HBV infection is essential to determining subsequent therapeutic response, including the development of drug-resistant infection..... monitoring of HBV response by measurement of HBV load was done a median of <1 time during the first year of therapy, with no consistent timing of HBV load measurement. Lack of overall follow-up care was not the cause of fewer HBV load measurements. The percentage of patients who had an HBV load or HBeAg and anti-HBe measurements increased somewhat over the course of the study, from <10% of patients at study initiation to about 50% of patients in the later years of the study. These data suggest that HIV care providers did not understand the need for and use of viral and serologic markers to monitor HBV treatment response. Because monitoring aminotransferase levels is part of HIV care, it is unclear if these tests were being used as surrogate markers of HBV response. The goal of therapy, which is suppression of HBV replication, may have been met in some patients, but it was not systematically determined.... Among the entire cohort of HIV-HBV-coinfected patients, 130 (36%) of 357 received >1 abdominal ultrasound examinations. Despite this, the number of positive findings was remarkably high: 50% of the abdominal ultrasounds that were performed had an abnormal finding, with liver cirrhosis being the most common abnormality. One hundred and sixty-one patients (45% of the entire cohort) had an AFP measurement at any time point in the study.....
....Adequate monitoring of HBV disease activity in coinfected patients helps identify viral seroconversion, the development of drug-resistant mutant strains, and hepatic decompensation. On the basis of our findings, we propose modifying HIV treatment guidelines to focus not only on the selection of drugs but also on monitoring for therapeutic response in patients with HIV-HBV coinfection."
ABSTRACT
Background. The common occurrence of hepatitis B virus (HBV) infection in patients who carry the human immunodeficiency virus (HIV) demands that both viruses be recognized, evaluated, and treated when appropriate.
Methods. We identified 357 HIV- and hepatitis B surface antigen-positive patients who underwent testing from 1999 to 2003; 155 patients who were new to our clinic and who initiated therapy for HIV and HBV coinfection were considered for inclusion in the study. The frequency of HIV testing (to determine HIV load and CD4+ cell count) performed during the first year of therapy was compared with the frequency of HBV measurements (to determine hepatitis B e antigen, antibody to hepatitis B e antigen, and HBV load), abdominal ultrasound examination, and measurement of levels of a-fetoprotein in serum.
Results.
HBV load data were obtained for only 16% of patients before initiation of antiretroviral therapy (ART), whereas HIV load was determined for 99% of patients before initiation of ART. The total number of HIV load measurements obtained during the first year after ART initiation was 497 (median number of HIV load measurements per patient, 3.0), compared with 85 measurements of HBV load (median number of HBV load measurements per patient, <1; P < .001). The percentage of patients who received any level of HBV monitoring (i.e., tests to determine hepatitis B e antigen, antibody to hepatitis B e antigen, and HBV load) after ART initiation increased from 7% in 1999 to 52% in 2001 (P < .001), whereas the percentage of patients who underwent HIV load testing remained at 80%-90% during the same period.
Conclusions. Health care providers treating patients with HIV infection during the period 1999-2003 infrequently monitored HBV response in coinfected patients, but they systematically monitored HIV response after ART initiation. Improved physician adherence to guidelines that better delineate HBV treatment and monitoring for patients with HIV-HBV coinfection is needed.
In the United States, up to 10% of all HIV-infected persons are coinfected with hepatitis B virus (HBV) [1-3]. As HIV-infected patients live longer as a result of antiretroviral therapy (ART), HBV infection has increasingly gained importance as a cause of death in such patients [2-5]. When patients receive drugs with dual activity (e.g., lamivudine, tenofovir disoproxil fumarate, and emtricitabine), they may experience acute flares of hepatitis resulting from immune reconstitution, drug withdrawal, the development of HBV drug-resistance mutations, or all 3 causes [6-12]. The increased risk of hepatocellular carcinoma (HCC) [13] and the emergence of HBV drug-resistant mutations necessitates careful monitoring of coinfected patients after antiviral therapy is begun.
Over the last 10 years, HBV therapy has improved considerably. Suppression of HBV replication using nucleoside and nucleotide analogues results in histologic improvement, slower disease progression, and increased long-term survival [14]. Effective case management for patients receiving drugs active against HBV requires monitoring HBV disease markers, aminotransferase levels, hepatitis B e antigen (HBeAg) status, and, in particular, HBV load to determine if therapy is needed and to assess response to treatment [15-17]. The goal of HBV therapy continues to be suppression of HBV replication and, if possible, HBeAg clearance [16, 18]. In addition, HBV management guidelines recommend screening for cirrhosis with imaging, -fetoprotein (AFP) testing, and liver biopsy, if appropriate [16]. Previous HIV treatment guidelines [19] have outlined both when HIV therapy should be initiated and when HIV loads should be measured, and current guidelines [20] recommend agents for treating patient with HIV-HBV coinfection but do not suggest how to monitor HBV response. To date, no studies have focused on how HIV care providers evaluate and monitor HIV-HBV-coinfected patients. Thus, the goal of our study was to determine the level of care for HBV infection that was being provided to patients with HIV infection by their health care providers.
PATIENTS AND METHODS
Of 357 hepatitis B surface antigen (HBsAg)-positive patients seen from 1999 to 2003 at Parkland Health and Hospital System outpatient HIV clinic (Dallas, TX), we identified 155 who were new to the HIV clinic and who had initiated ART after 1999; these patients formed the basis of our study. Of the 357 patients seen from 1999 to 2003, 124 (35%) were excluded from the study because they initiated ART before 1999, 45 (13%) were excluded because they had test results positive for HBsAg after ART initiation, 32 (9%) were excluded because they had no HIV or HBV therapy documented, and 1 was excluded because HBsAg seroconversion had occurred prior to ART initiation. In the group of patients who did not have a record of ART initiation, most had high CD4+ cell counts and/or did not return for follow-up visits and were never assessed fully for severity of liver disease. For the 155 patients in the cohort, baseline data collected included patient demographic data, serum aminotransferase levels, CD4+ cell counts, HIV load, HCV antibody (anti-HCV) levels, hepatitis delta antibody (anti-HDV) levels, risk factors for HIV acquisition, and HIV and/or HBV therapy (including the date of therapy initiation). We determined the frequency of HBV testing (i.e., testing to determine HBeAg, antibody to HBeAg [anti-HBe], and HBV load) from 90 days prior to baseline to 1 year after initiation of therapy. In addition, the frequency of HIV load and CD4+ cell count measurements was determined during the first year after initiation of therapy. Patients were grouped by frequency of HIV load measurements into those with 0-1, 2-3, and >4 virus load measurements during the first year of therapy. This served as a surrogate for the number of follow-up visits. To examine overall monitoring for HBV infection, we determined the frequency of HBV testing (i.e., testing to determine HBeAg, anti-HBe, and HBV load) and screening for HCC (i.e., ultrasound or AFP measurements). In addition, we examined pharmacy data for patients with CD4+ cell counts <200 cells/uL to determine whether this group was administered pneumocystis pneumonia prophylaxis, as recommended by HIV practice guidelines [21], in a timely fashion (i.e., within 2 months) after ART initiation. We determined the time intervals from initiation of ART to first HBV DNA quantification assay or HBeAg and anti-HBe test and from initiation of treatment to first measurement of HIV RNA level and CD4+ cell count for every year of the study. The University of Texas Southwestern Medical Center's institutional review board approved the study.
Jonckheere-Terpstra analysis was used to test ordered differences among start year and HBV testing and HIV load testing, as well as to test HIV load groups and HBV and HCC testing during the first year of therapy. The X2 test or Fisher's exact test were used to compare the occurrence of HBV and HCC testing for patients compared with baseline characteristics. Comparisons of time to and frequency of determination of HBV load and HIV load were analyzed by the sign test for paired data. A P value of <.05 was considered to be statistically significant.
RESULTS
Use of drugs active against HBV. The patients' baseline demographic and clinical data are shown in table 1; the majority of patients were male, and 52% were black. No patients were identified as Asian. Twenty-one (14%) of 154 patients had test results positive for anti-HCV, whereas only 1 (2.5%) of 41 patients had test results positive for anti-HDV. Ninety-two percent of patients received nucleosides and/or nucleotides that were active against HBV; 88% of patients received lamivudine.
Initial HIV and HBV testing. All but 1 of the 155 patients had at least 1 HIV load measurement, and all 155 patients had at least 1 CD4+ cell count assessment before initiation of ART. However, only 25 (16%) received testing to determine HBV load and HBeAg and anti-HBe levels before initiation of ART; 27 (17%) of 155 had an ultrasound examination of the liver, and 51 (26%) of 155 had an AFP test performed prior to or within 6 months after ART initiation.
Monitoring HIV and HBV infection during the first year of ART. The total number of tests to determine HIV load performed during the first year of therapy was 497 (median number of tests per patient, 3 per year), whereas only 85 tests were performed to determine HBV load during the same time period in the same patient group (median number of tests per patient, <1 per year; P < .001, by the sign test). Failure to measure HIV loads at any time during the first year after ART initiation was observed in 20 (13%) of the patients, whereas HBV load was not measured for 104 (67%) during the same time period. Because lack of follow-up may have influenced whether HBV tests were performed, we grouped patients into those who had 0-1 (35 patients), 2-3 (49 patients), and >4 (71 patients) HIV load tests performed during their first year of treatment, as an approximation of the number of follow-up visits. No relationship was found between frequency of HIV load or CD4+ cell count measurements and any HBV testing (data not shown). More frequent follow-up visits were not associated with more frequent HBV assessments: only one-third of patients had HBV tests performed at any time in the group that had >4 HIV load measurements obtained. However, a higher proportion of AFP measurements and a trend towards more ultrasound examinations was observed in the group who had more HIV load measurements.
Improvement in monitoring of HBV infection during the study. The proportion of patients receiving any HIV load testing during the first year of ART remained constant from 1999 to 2003. However, although only 7%-13% of patients who started receiving HIV medications in 1999 or 2000 had some form of HBV testing during the year after ART initiation (figure 1), a significantly higher proportion of those initiating ART from 2001 to 2003 received HBV testing (43%-52%; P<.001, by Jonckheere-Terpstra analysis). Despite this improvement, by 2003, nearly 50% of the patients still did not receive any monitoring for HBV response, whereas 80%-90% of these patients received HIV monitoring during the same time period.
Time to initial HBV load measurement. After ART initiation, the time to first HBV load measurement, compared with the time to first HIV load measurement, was greatly delayed, both in 1999 (median time to first measurement, 1067 days vs. 46 days; P < .001, by the sign test) and in 2000 (median time to first measurement, 664 days vs. 59 days; P = .002, by the sign test). However, this delay decreased considerably by 2001, and by 2003, at least those who received HBV load measurements obtained them near the time of ART initiation.
HBV evaluations in patients with and patients without AIDS. No difference in overall HBV evaluation rates was observed between patients with CD4+ cell counts <200 cells/uL and patients without AIDS. We found that 90 (97%) of the patients with CD4+ cell counts <200 cells/uL were given pneumocystis pneumonia prophylaxis at the time of ART initiation.
Outcomes and monitoring for complications. Monitoring for complications of HBV infection (e.g., for development of HCC) was done infrequently. Among the entire cohort of HIV-HBV-coinfected patients, 130 (36%) of 357 received >1 abdominal ultrasound examinations. Despite this, the number of positive findings was remarkably high: 50% of the abdominal ultrasounds that were performed had an abnormal finding, with liver cirrhosis being the most common abnormality. One hundred and sixty-one patients (45% of the entire cohort) had an AFP measurement at any time point in the study.
DISCUSSION
Our data strongly indicate that, during the period from 1999 to 2003, HIV care providers frequently failed to initially evaluate or to subsequently monitor HBV infection status in coinfected patients despite careful follow-up for HIV infection. Although every patient is tested for HBsAg, anti-HCV, and antibodies for hepatitis A, only 16% had had an HBV load measured to assess the status of disease and the possible need for treatment of HBV infection prior to initiation of ART. However, 92% of this patient group received ART regimens that contained .1 HBV-active nucleosides and/or nucleotides. Assessing baseline viral and serologic markers of HBV infection is essential to determining subsequent therapeutic response, including the development of drug-resistant infection. Because we only evaluated the initial regimen, it is unclear if follow-up adjustments were made to ART because of the development of drug-resistant HBV infection.
After the initiation of ART, adequate monitoring of HIV response was observed, with a median of 3 measurements of HIV load and CD4+ cell count performed each year. HIV load measurements were consistently performed 6-8 weeks after ART initiation in all years of the study, and HIV loads were measured for 80%-90% of patients during each year. On the other hand, monitoring of HBV response by measurement of HBV load was done a median of <1 time during the first year of therapy, with no consistent timing of HBV load measurement. Lack of overall follow-up care was not the cause of fewer HBV load measurements. The percentage of patients who had an HBV load or HBeAg and anti-HBe measurements increased somewhat over the course of the study, from <10% of patients at study initiation to about 50% of patients in the later years of the study.
These data suggest that HIV care providers did not understand the need for and use of viral and serologic markers to monitor HBV treatment response. Because monitoring aminotransferase levels is part of HIV care, it is unclear if these tests were being used as surrogate markers of HBV response. The goal of therapy, which is suppression of HBV replication, may have been met in some patients, but it was not systematically determined.
Although HBV practice guidelines [16, 17] suggest regular monitoring of HBV loads, no similar recommendations for determining HBV response in HIV-HBV-coinfected patients are available in either current or former guidelines [19, 20, 22, 23]. Guidelines on monitoring HBV infection after ART initiation were limited during the period 1999-2003, which may explain the lack of monitoring for therapeutic response and clinical complications. Several guidelines [24-28], published after the time period of this study, do focus on the use of drugs and serologic and virologic markers of response; however, only 1 guideline [25] indicates how frequently this should occur. We have now implemented a detailed guideline in our clinic to standardize the management of HIV-HBV-coinfected patients (figure 2).
Our study was limited by its retrospective nature, its small sample size, and the fact that it was based on clinical practice in a single health care center. Data on the frequency of HBV load, HBeAg, and anti-HBe measurements, ultrasound examinations, and AFP tests were based on the tests that were completed. It was not possible to review medical records to determine how often providers ordered tests that were never performed. In addition, radiographic tests are less likely to be completed than serologic tests.
In summary, HIV care providers often failed to monitor HBV response in their overall management of HIV-HBV-coinfected patients prior to and after initiation of ART. Adequate monitoring of HBV disease activity in coinfected patients helps identify viral seroconversion, the development of drug-resistant mutant strains, and hepatic decompensation. On the basis of our findings, we propose modifying HIV treatment guidelines to focus not only on the selection of drugs but also on monitoring for therapeutic response in patients with HIV-HBV coinfection.
|
|
| |
| |
|
|
|
