| |
Utilization of Surveillance for Hepatocellular Carcinoma Among Hepatitis C Virus-Infected Veterans in the United States: 'HCC surveillance low among cirrhotics despite recommendations for HCC surveillance in these high-risk patients'
|
| |
| |
Download the PDF here
Annals of Internal Medicine Jan 18 2011
"routine HCC surveillance with either an AFP test or ultrasonography is low among HCV-infected patients with cirrhosis, despite recommendations for HCC surveillance in these high-risk patients. Future studies are needed to evaluate the knowledge, attitudes, and barriers for HCC surveillance and to develop appropriate, targeted interventions to increase the dissemination of this practice."
"Hepatocellular carcinoma (HCC) is the fastest rising cause of cancer-related deaths in the United States ....Previous studies reported very low rates of HCC surveillance before diagnosis among patients with HCC, even in the presence of a recorded cirrhosis diagnosis (22-24). No studies have examined the frequency or patterns of surveillance in HCV-infected patients with cirrhosis, a scenario that more likely reflects real-life practice settings in the United States. Current practice guidelines recommend screening for hepatocellular carcinoma (HCC) in patients with cirrhosis. The evaluation of data in a Veterans Affairs database showed that routine, annual screening for HCC with either serum α-fetoprotein measurement or abdominal ultrasonography was done in only 12% of veterans with cirrhosis. Testing was done inconsistently in 58.5% and not at all in 29.5% of patients with cirrhosis. This study could not determine whether missing screening was due to physicians' failure to recommend tests or patients' failure to adhere to testing. Efforts are needed to improve screening for HCC in at-risk patients."
-----------------------
Surveillance for hepatocellular carcinoma (HCC) in patients with cirrhosis is recommended but 17% received regular testing for HCC(liver cancer)
Surveillance for hepatocellular carcinoma (HCC) in patients with ...
www.natap.org/2011/HCV/021411_01.htm "in the United States HCC surveillance is not applied as widely as it is in many European and Far Eastern countries".
------------------------
Abstract
Background: Surveillance for hepatocellular carcinoma (HCC) is recommended for patients with hepatitis C virus (HCV) infection and cirrhosis. However, whether surveillance is being done as recommended is unknown.
Objective: To examine the prevalence and determinants of HCC surveillance among HCV-infected patients with cirrhosis in Veterans Affairs (VA) health care facilities in the United States.
Design: Retrospective cohort study of HCV-infected patients using data obtained from the national VA Hepatitis C Clinical Case Registry.
Setting: 128 VA medical centers.
Patients: HCV-infected patients with cirrhosis diagnosed between fiscal years 1998 and 2005.
Measurements: Abdominal ultrasonography and measurement of α-fetoprotein for HCC surveillance were identified from administrative data by using a previously validated algorithm. Patients were categorized as having routine (tests done during at least 2 consecutive years in the 4 years after cirrhosis diagnosis), inconsistent (at least 1 test, but not routine), or no surveillance in the 4 years after cirrhosis diagnosis. Predictors of surveillance were identified by using hierarchical random-effects regression.
Results: 126 670 patients with HCV were identified; 13 002 (10.1%) had cirrhosis. Approximately 42.0% of patients with cirrhosis received 1 or more HCC surveillance tests within the first year after the cirrhosis index date; however, a decline in receipt of surveillance was observed in the following 2 to 4 years. Among patients with cirrhosis and at least 2 years of follow-up, routine surveillance occurred in 12.0%, inconsistent surveillance in 58.5%, and no surveillance in 29.5%. Lower medical and psychological comorbid conditions, presence of varices, and the absence of decompensated liver disease were associated with a higher likelihood of receiving routine surveillance.
Limitations: Hepatocellular carcinoma surveillance tests were indirectly identified from registry data. Physician recommendations could not be captured.
Conclusion: Few HCV-infected veterans with cirrhosis received routine HCC surveillance. New strategies are needed to improve the implementation of HCC surveillance in clinical practice.
Hepatocellular carcinoma (HCC) is the fastest rising cause of cancer-related deaths in the United States (1). This increase is mostly attributable to an increase in hepatitis C virus (HCV)-related HCC (2-5). Survival with HCC is generally very poor (overall 5-year survival is <5%), except when patients receive potentially curative therapy in the form of a liver transplant, surgical resection, or tumor ablation. In these patients, a considerable improvement in survival has been observed (5-year survival ranges from 40% to 70%) (6).
Although several treatment options for HCC now exist, the eligibility of patients to receive these treatments, as well as the effectiveness of these treatments, diminishes with more advanced disease. Therefore, practice guidelines recommend HCC surveillance in high-risk groups (for example, HCV-infected patients with cirrhosis) in order to detect earlier-stage HCC and ultimately increase receipt of treatment and improve survival (7, 8).
Abdominal ultrasonography and measurement of serum α-fetoprotein (AFP) are the 2 most commonly recommended tests for HCC surveillance (7-13). Although no randomized, controlled trials of HCC surveillance have been done in HCV-infected patients, 1 randomized, placebo-controlled trial in hepatitis B carriers, as well as several observational cohort and case-control studies in patients with HCV, hepatitis B, and alcoholic cirrhosis, have shown that HCC surveillance is associated with earlier HCC diagnosis, greater use of potentially curative therapy, and a significant reduction in cancer-specific mortality compared with patients with symptomatic HCC (14-21).
It is unclear how often HCC surveillance is done among at-risk patients in clinical practice. Previous studies reported very low rates of HCC surveillance before diagnosis among patients with HCC, even in the presence of a recorded cirrhosis diagnosis (22-24). No studies have examined the frequency or patterns of surveillance in HCV-infected patients with cirrhosis, a scenario that more likely reflects real-life practice settings in the United States.
The Veterans Affairs (VA) health care system is the largest integrated health care system in the United States and has a disproportionate number of patients with HCV. Moreover, the VA is a semiclosed system with a relatively stable patient population, which makes it a suitable setting for examining how often surveillance is done and variations in HCC surveillance practices. We therefore conducted a retrospective cohort study of all eligible HCV-infected patients with cirrhosis to identify patterns and determinants of HCC surveillance. We also examined a cohort of HCV-infected patients without cirrhosis, a group in which guidelines do not recommend surveillance.
Results
We identified 126 670 patients with an HCV index date from 1 October 1997 to 30 September 2005 (fiscal years 1998 to 2005) who fulfilled the inclusion and exclusion criteria. Of these patients, 13 002 (10.2%) had received a diagnosis of cirrhosis (Table 1). Patients were included in the prevalence of surveillance estimates presented in Table 2 only if they had complete follow-up in the indicated periods, as defined in the Methods section. We excluded 2076 patients with cirrhosis who had less than 2 years of complete follow-up during the 4 years after the index date for the routine, inconsistent, and no surveillance estimates. We found that only 12.0% of patients with cirrhosis received routine surveillance, 58.5% received inconsistent surveillance in at least 2 consecutive years in the 4 years after the cirrhosis index date, and 29.5% did not receive any surveillance. To calculate the 1-, 2-, 3-, and 4-year prevalence of surveillance estimates in Table 2, we excluded 1557, 2076, 1843, and 2249 patients with cirrhosis, respectively, because of incomplete follow-up during the indicated periods. Approximately 42.0% (n = 4809) of patients with cirrhosis received at least 1 AFP test or ultrasonography for HCC surveillance during the 1 year after the index date; of these, 34.4% received an AFP test, 20.7% received ultrasonography, and 44.9% received both tests. The proportions of patients with cirrhosis who received at least 1 AFP test or ultrasonography for surveillance remained steady (about 36%) during years 2, 3, and 4 after the cirrhosis index date.
In a sensitivity analysis of all patients with cirrhosis, regardless of whether complete follow-up information was available, receipt of an AFP test or ultrasonography for HCC surveillance in year 2 decreased to 33.2% (3797 of 11 445), 31.3% (3581 of 11 445) in year 3, and 28.1% (3219 of 11 445) in year 4. These estimates included HCC surveillance tests for patients who were previously excluded because of incomplete follow-up. Patients with 4 years of complete follow-up (n = 5277) were significantly less likely to drink alcohol than patients with shorter follow-up periods (n = 6168), respectively (42.8% vs. 48.7%; P < 0.001). No significant differences in follow-up were observed regarding the use of cocaine or cannabis.
Patients with cirrhosis who received routine surveillance had a lower Model for End-Stage Liver Disease score and were less likely to have several specific comorbid conditions (ascites, coronary artery disease, chronic obstructive pulmonary disease, diabetes, psychosis, alcohol use, and other substance use) than were those who had not received surveillance and were more likely to receive a diagnosis of cirrhosis during more recent years (Table 3). These factors remained statistically significant in a multilevel model in which patients were grouped hierarchically within VA providers and VA facilities (Table 4).
In addition to the temporal trends of HCC surveillance based on the cirrhosis index date, we examined annual changes in receipt of HCC surveillance in patients with cirrhosis, regardless of time of diagnosis. A trend toward increasing receipt of routine surveillance was observed from 6.9% in 1998 to 19.9% in 2005.
Among 113 668 patients without cirrhosis, approximately 29.7% received at least 1 AFP test or ultrasonography for HCC surveillance during the year after the HCV index date, and only 3.3% of patients received routine surveillance (Table 2). Among these patients, 32.8% received an AFP test, 29.9% received ultrasonography, and 37.3% received both tests.
In a sensitivity analysis examining the frequency of all AFP test, ultrasonographies, and computed tomographies, regardless of the purpose, the proportion of patients who received at least one of these tests remained higher among patients with cirrhosis (72.3% in the first year, decreasing to 57.0% in the fourth year after the cirrhosis index date) (Table 2).
Among the 39 089 Medicare-eligible patients in the study cohort, less than 1.0% of patients had an AFP test (115 at year 1, 106 at year 2, and 66 at year 3) or ultrasonography (234 at year 1, 232 at year 2, and 137 at year 3) identified in Medicare claims only in the first 3 years after the HCV index date.
Discussion
In this study of care within the VA health care system, most (88%) HCV-infected patients with cirrhosis did not receive routine HCC surveillance, as recommended by guidelines. Approximately 42% of patients with cirrhosis received at least 1 surveillance test within the first year after the cirrhosis index date; however, receipt of HCC surveillance considerably decreased in the following 2 to 3 years. Most HCV-infected patients with cirrhosis received sporadic or no HCC surveillance.
Our primary analyses focused on routine surveillance using AFP tests or ultrasonography, as identified indirectly using a validated algorithm for HCC surveillance (27). However, it is possible that ultrasonography or computed tomography may be done to diagnose other conditions, such as gallbladder or pancreas disorders. Although these tests do not count as surveillance tests, they can affect future decisions about performing additional HCC surveillance tests. We found that at least one of these tests was done in 72.3% of patients with cirrhosis (of which computed tomography accounted for only 1.2%) in the first year after the cirrhosis index date and in each subsequent year, slightly more than half of patients received any of these tests. These utilization rates probably overestimate the true prevalence of HCC surveillance. Furthermore, performing a single or an irregular surveillance test is not likely to be very useful. Therefore, we believe that our primary analyses of routine receipt of HCC surveillance tests in patients with cirrhosis present the most relevant findings of the study.
Implementation of HCC surveillance guidelines is low in clinical practice. Several factors may contribute to this observation, including the need for repeated surveillance during relatively short periods, potential difficulty in following up with patients after a positive or an equivocal surveillance test, the somewhat complicated diagnostic evaluation for HCC, and the limited availability of liver transplantation centers to refer patients who receive a diagnosis of HCC. Future studies are needed to examine the effect of these factors on the utilization of HCC surveillance in clinical practice. The low estimates of HCC surveillance that we report sharply contrast the findings of a 1998 survey, in which 84% of hepatologists claimed that they did routine surveillance for HCC (32). It is possible that the survey estimates may be inflated because of recall bias or that they reflect only a small segment of specialized providers that do not represent most physicians involved in the care of patients with cirrhosis.
Greater diffusion of practice guidelines probably accounts for some of the increase in routine surveillance in patients with more recent diagnoses of cirrhosis observed in this study. Similar findings were observed in our previous study among Medicare patients conducted by using the Surveillance, Epidemiology, and End Results and Medicare databases (merged to become 1 database for research purposes) (23). Although HCC surveillance recommendations were available before the beginning of our study in 1998, the main international guideline recommending HCC surveillance was released by the European Association for the Study of the Liver in 2001 (7). Nevertheless, the rate of increase is slow, and the overall implementation seems to be very inadequate. We recommend planning active interventions to improve HCC surveillance.
Current practice guidelines do not recommend HCC surveillance in HCV-infected patients without cirrhosis. However, approximately 29.7% of HCV-infected patients without cirrhosis had at least 1 surveillance test in the first year after diagnosis, and 3.3% had routine surveillance. It is possible that some of these patients had cirrhosis that was missed by our study definition. Although surveillance in these patients may not be inappropriate or wasteful, this provides evidence of confusion and poor implementation of guidelines, especially in the context of low utilization of HCC surveillance in the group at highest risk (that is, patients with cirrhosis).
We found that several patient-related factors, including presence of comorbid conditions, advanced liver disease, and alcohol use, were associated with a lower likelihood of receiving routine HCC surveillance. The presence of severe comorbid conditions may reduce the likelihood of receiving potentially curative therapy if HCC should develop, thereby diminishing physician enthusiasm for surveillance. We could not accurately ascertain the severity or reversibility of comorbid conditions in our study. However, with the emergence of efficacious palliative treatments, such as ablation, transarterial chemoembolization, and sorafenib, excluding patients with moderate or controlled comorbid conditions from HCC surveillance may not be justified (8, 33, 34). Patients who received routine surveillance were slightly more likely to have mild to moderate liver disease; this is to be expected because HCC surveillance may be futile in patients with advanced cirrhosis not listed for transplantation (35). Finally, patients who drank alcohol were less likely to receive routine HCC surveillance and also were less likely to have 4 years of complete follow-up after the cirrhosis index date, thus suggesting that these patients were less likely to be engaged in regular health care activities, including receipt of HCC surveillance.
Patients who use the VA health care system may receive care outside the VA using supplemental insurance or, most likely in our sample, Medicare benefits. However, less than 1% of all Medicare-eligible patients in our study had an AFP test or ultrasonography identified in Medicare claims files only. Given the large number of patients in our study cohort, the few tests identified in Medicare claims would not have substantially changed our results.
Our findings should be interpreted within our study's limitations. Hepatocellular carcinoma surveillance tests could not be directly identified from administrative data (29). Therefore, we developed and validated an algorithm with good predictive values to identify both AFP tests and ultrasonography done for surveillance purposes. Although the algorithm for ultrasonography less accurately identifies HCC on a surveillance test (27), the total number of ultrasonographies was minimal compared with the total number of AFP tests. Therefore, misclassification of ultrasonography would have only a small effect on the overall estimates of HCC surveillance. Finally, we were unable to capture physician recommendations to perform surveillance or patient adherence to these recommendations.
To our knowledge, our study is the largest to date on HCC surveillance in HCV-infected patients with cirrhosis. Our study includes other strengths: several years of complete follow-up, inclusion of data from recent years reflecting contemporary practice, accurate definitions of cirrhosis validated by previous chart reviews using our study database, and examination of a range of variables that may affect the utilization of HCC surveillance.
In conclusion, routine HCC surveillance with either an AFP test or ultrasonography is low among HCV-infected patients with cirrhosis, despite recommendations for HCC surveillance in these high-risk patients. Future studies are needed to evaluate the knowledge, attitudes, and barriers for HCC surveillance and to develop appropriate, targeted interventions to increase the dissemination of this practice.
|
|
| |
| |
|
|
|
