| |
HCC (Liver Cancer) Deaths Doubled & Hospital Services Costs Increase Significantly
|
| |
| |
Gastroenterology
August 2005
.....rising burden of HCC....
"Mortality and Hospital Utilization for Hepatocellular Carcinoma in the United States"
.....the burden of HCC increased substantially in terms of mortality and hospital service utilization... It is important to point out the progressive increase in hospital service utilization seen in this analysis. For example, there was a 40% increase in hospital charges between 1998 and 2000 alone....
.....To the degree that the prevalence of these conditions (chronic liver disease because of hepatitis B and C and possibly NAFLD) is expected to persist, if not increase, the burden of HCC is likely to continue in the future.
Closer epidemiologic surveillance is needed to understand better the causation of HCC at the population level, and judicious use of that knowledge to implement primary and secondary prevention measures is urgently warranted.
..... The total number of HCC deaths increased more than 2-fold, from 3175 in 1980 to 6984 in 1998...... In the present analysis, we found a 2.2-fold increase in the number of deaths and 67% increase in mortality rate between 1980 and 1998.... Although the number of HCC deaths increased in all categories, the proportion with viral hepatitis increased from <1% to 15% over time..... Recent data from Taiwan indicate that antiviral therapy might be able to reduce the occurrence of HCC among those with advanced chronic hepatitis B....
..... There were estimated 23,731 HCC hospitalizations nationwide in 2000, approximately 76% higher than 1988 (13,171 hospitalizations). The distribution of liver diagnosis in 2000 included viral hepatitis (28%)..... One potential scenario may be nonalcoholic fatty liver disease, which may remain silent even when cirrhosis or HCC develops
This analysis was able to gain a partial insight into the underlying cause of increasing frequency of HCC. For example, in the NIS data, increasing proportions of HCC hospitalizations were associated with hepatitis B or C, which accounted for 28% of all HCC hospitalizations and almost 20% of all in-hospital deaths in 2000. A similar trend was noted in the mortality data, although ascertainment of underlying liver disease was much less complete. There is little doubt that part of this observed increase represents a detection bias in that testing for hepatitis C became available only in 1990, and many deaths related to hepatitis C had not been recognized as such before the early 1990s.12 Although the magnitude of this ascertainment problem is difficult to gauge, it appears likely that increasing numbers of people of long-standing hepatitis C infection is partly responsible for this trend.
Authors: W. Ray Kim_, Gregory J. Gores_, Joanne T. Benson_, Terry M. Therneau, L. Joseph Melton III
Background & Aims: The incidence of hepatocellular carcinoma (HCC) has been increasing in the United States. Although resource-intensive treatment modalities have been increasingly applied, these patients still have poor survival. We examined 2 nationally representative databases, the Multiple Cause of Death file and the Nationwide Inpatient Sample database, to examine trends in mortality and hospital service utilization related to HCC.
Methods: In both databases, a priori criteria were used to identify cases of HCC. All other available diagnostic fields were examined to characterize coexistent liver disease. Age-, sex-, and race-specific mortality from HCC was calculated, and temporal changes in mortality rates were evaluated using the multivariable Poisson model. Hospital service utilization was estimated based on length of stay, total hospitalization charges, and principal procedures.
Results:
The age-, sex-, and race-specific mortality from HCC increased from 1.54 to 2.58 per 100,000 per year between 1980 and 1998.
Male sex, African and Asian race, and increasing age were also associated with higher mortality.
The estimated total charge for HCC hospitalizations nationwide increased from $241 million in 1988 to $509 million in 2000 after inflation adjustment.
Commonly employed procedures in 2000 included angiography/embolization, resection, local ablative therapy, and liver transplantation.
Conclusions: In the recent past, mortality and hospital service utilization related to HCC increased substantially. Closer epidemiologic surveillance to understand causation of HCC at the population level and to help implement primary and secondary prevention is urgently warranted.
BACKGROUND
According to the World Health Organization, hepatocellular carcinoma (HCC) is one of the most common malignancies in the world, with approximately 398,000 incident cases in 2000.1 It is also a highly lethal malignancy, resulting in nearly 384,000 deaths that same year.2 In the United States, the incidence of HCC, based on the Surveillance, Epidemiology, and End Results (SEER) program, is reportedly increasing,3 and, because of the poor survival of these patients, the mortality associated with HCC is expected to have risen in parallel with the rise in incidence.4 The vast majority of HCC arises in patients with underlying chronic liver disease, most commonly cirrhosis,5 and the rise in the incidence and mortality from HCC in the United States has been attributed to increasing numbers of Americans with cirrhosis from long-standing chronic hepatitis C virus (HCV) infection.6 However, the extent to which HCV infection accounts for the trend of HCC epidemiology remains uncertain because there are other epidemiologic trends in liver disease that may also have contributed to the changing epidemiology of HCC. For example, the burden of hepatitis B virus (HBV) infection has been increasing because of recent influx of immigrants from endemic countries.7 In addition, the prevalence of nonalcoholic fatty liver disease (NAFLD) is also believed to be increasing as a result of increasing obesity among Americans.8,9
During the 1990s, several new therapeutic approaches have become available for patients with HCC. These include locally ablative interventions such as percutaneous ethanol injection or radio-frequency ablation as well as liver transplantation.10 Although the impact of these interventions on the population-wide mortality from HCC remains to be determined, some of these interventions are quite resource intensive (eg, liver transplantation). Data on the economic impact of HCC in recent years have not been closely examined.
In this work, we hypothesize that there have been increasing trends in mortality and hospital service utilization from HCC in the United States in the recent past. To examine this overall hypothesis, we use detailed death registry data to assess mortality from HCC and nationally representative health care utilization data to estimate hospital service utilization in the United States. In addition to evaluating time trends, we seek to determine demographic and clinical characteristics that may underlie the increase in HCC and to examine the impact of recent advances in HCC management on hospital service utilization on a nationwide scale.
Data Sources
Two databases were used to analyze trends in mortality and hospital service utilization for HCC. First, data on mortality were obtained from the Multiple Cause of Death files, which represent a compilation of all death certificates (approximately 2 million annually) issued in the United States (often referred to as the "Death Registry"). Data on even calendar years between 1980 and 1998 were obtained from the National Center for Health Statistics (Hyattsville, MD). Available data elements included age, sex, race, and place of birth and death, and all listed (immediate, contributing, and underlying) causes of death on the death certificate.
Second, hospital service utilization for HCC was estimated based on the Nationwide Inpatient Sample (NIS) of the Healthcare Utilization Project database available from the Agency for Healthcare Research and Quality (AHRQ; Rockville, MD). Although the death registry includes all deaths that occurred in the United States, the NIS data represent a 20% sample (6-7.5 million discharge records per annum) of nonfederal, acute care hospitals, which altogether house 85% of hospital beds and account for 94% of hospitalizations in the United States. To ensure that NIS is representative of the United States as a whole, AHRQ uses a sophisticated stratified sampling method, based on ownership control, bed size, teaching status, urban/rural location, and region of the United States. Discharge weights, sampling frame, and strata are provided for individual records, which allows calculation of nationwide estimates along with 95% confidence intervals (95% Cl). NIS data from all available calendar years between 1988 and 2000 were obtained. Available data elements included age, sex, race of the patient; length of stay; procedures; hospital charges; discharge diagnoses; and disposition including in-hospital death.
Results
Deaths from HCC
In the Multiple Cause of Death file, there were a total of 80,874 records of potential HCC deaths in the 10 even calendar years between 1980 and 1998. Of those, 41,439 (51%) had the ICD-9 code of 155.0, whereas the remaining 39,435 (49%) were coded to 155.2; 6910 (18%) of this latter group also included another diagnosis of liver disease and, thus, met the predefined inclusion criteria for HCC death. Therefore, a total of 48,349 deaths were included in the analysis as HCC. Of these, there were 33,487 (69%) men and 14,862 women. The mean age at death was 61.4 years for men and 65.3 for women. Approximately 16% of the decedents were born outside the United States. A large majority (78%) of the decedents were of white race, followed by African (14%), and other (8%), which included Asian, Pacific Islander, Native American, and Alaskan. Overall, only approximately 47% of HCC death records included an underlying liver disease diagnosis, such as alcoholic liver disease (3%), viral hepatitis (6%), and cirrhosis with no other specific diagnosis (29%). Disease frequencies were similar by sex except that alcoholic liver disease was more common in men (4%) than in women (1%).
The total number of HCC deaths increased more than 2-fold, from 3175 in 1980 to 6984 in 1998. Although the number of HCC deaths increased in all categories, the proportion with viral hepatitis increased from <1% to 15% over time. The proportion of deaths with alcoholic liver disease (3%) and those without a mention of underlying liver disease (52%) remained constant.
In conjunction with the increased number of deaths, the age-, sex-, and race-adjusted mortality rate rose 67% between 1980 and 1998, from 1.54 (95% CI: 1.48-1.60) to 2.58 (95% CI: 2.52-2.64) per 100,000. The increase in men was from 2.31 (95% CI: 2.21-2.42) to 4.24 (95% CI: 4.12-4.36) per 100,000 and, in women, was from 0.96 (95% CI: 0.90-1.01) to 1.25 (95% CI: 1.19-1.30) per 100,000. Thus, although the death rate increased in both sexes, the increase was higher among men (84% between 1980 and 1998) than women (30%). HCC mortality was low under 40 years of age. Thereafter, the mortality rates increased progressively with age in both men and women as they did also among members of each race.
When each variable was considered individually, HCC mortality was 2.47-fold higher in men than in women; it was 2.51 times higher in "other" race and 1.37 times higher in Africans than in whites. Age and calendar year were fit linearly by 10-year increments. A 10-year increase in age was associated with 2.02-fold increase in mortality, and, for each decade in calendar year, HCC mortality rose by 39%. The effect of male sex, age, and race on HCC mortality remained strong. For example, on the average, HCC mortality was 1.98-fold higher among people of African race and 3.82 higher among people of other races. On top of these effects, there were some differences in the temporal trend in HCC mortality by race and sex. For example, the rise in HCC mortality was 34% per decade among white men and 20% among white women. In general, the rise in HCC mortality followed a steeper slope in men than in women. Although the rate of increase was similar among men of the 3 racial groups, it was highest among white women, followed by women of other races. Among women of African race, there was no significant rise in HCC mortality over time.
Hospital Service Utilization for HCC
There were estimated 23,731 HCC hospitalizations nationwide (95% CI: 20,473-25,873) in 2000, approximately 76% higher than 1988 (13,171 hospitalizations [95% CI: 11,891-14,451]. The distribution of liver diagnosis in 2000 included viral hepatitis (28%), alcoholic liver disease (14%), and other liver disease (16%). Concurrent liver disease diagnosis was absent in 43% of the discharge records.
The estimated nationwide total charges for HCC hospitalizations increased from $241 million in 1988 to $509 million in 2000 after inflation adjustment. The way hospitalizations happened to be sampled, there was a greater variance (thus wider confidence intervals in estimates) for years 1990 and 1992. However, it is likely that there was a steady increase in hospital service utilization. This increase was not only attributable to the number of hospitalizations but also to the intensity of resource utilization per hospitalization. For example, the average daily charges increased $2140 (95% CI: 1966-2315) to $4007 (95% CI: 3563-4453) after inflation adjustment, indicating more activities per hospitalization day. In contrast, the average length of stay decreased from 9.6 (95% CI: 8.9-10.3) days in 1988 to 6.8 (95% CI: 6.2-7.3) in 2000. This trend of shortened length of hospitalization and higher charges per day of hospitalization was not unique to HCC hospitalizations. For all hospitalizations recorded in NIS, the mean length of stay decreased from 6.3 days (95% CI: 6.2-6.4) to 4.6 days (95% CI: 4.5-4.7), whereas mean daily charges increased from $1910 (95% CI: 1861-1958) to $3470 (95% CI: 3136-3804) between 1988 and 2000.
Approximately 70% of hospitalizations for both calendar years listed a principal procedure. Of the records that included a principal procedure, percutaneous liver biopsy, para/thoracentesis, and chemotherapy were most commonly encountered for both years. Not surprisingly, therapeutic procedures were more common in 2000, including angiography/embolization (5.5%), resection (4.9%), and local ablative therapy (3.5%). There were 302 (1.8%) hospitalizations in 2000 in which liver transplantation was the principal procedure. In contrast, explorative surgery decreased from 7.2% in 1988 to 1.3% in 2000. Endoscopic procedures were commonly utilized for both calendar years.
In-hospital Death
In the Multiple Cause of Death File, HCC death occurred in a hospital in 54% of cases. In the Nationwide Inpatient Sample, 19% of HCC hospitalizations ended with the patient's death. Figure 5 shows that the number of in-hospital deaths from the Multiple Cause of Death file were consistent with those reported in the Nationwide Inpatient Sample between 1990 and 1998. For example, in the Multiple Cause of Death file, there were 3267 in-hospital deaths from HCC in 1998 compared with 3171 (95% CI: 2911-3431) estimated by the Nationwide Inpatient Sample. There was, in general, good agreement between estimates from the 2 databases with regard to age and sex characteristics. The demographic characteristics of decedents also appeared similar in the 2 data sets (data not shown).
Discussion
The rising incidence of HCC in the United States has been reported previously. According to the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute, the overall incidence of HCC increased by 114% between 1975 and 1998.3,4 Because survival of patients with HCC is usually limited, a corresponding increase may be predicted in mortality rates, but only limited data on HCC mortality are available. For example, a previous publication reported that mortality from HCC had indeed increased since early 1980s to more than 22,000 deaths annually during the first half of the 1990s.3,4 A recent statistic from the American Cancer Society, however, described a lower figure for 2003: Deaths from liver and intrahepatic bile duct cancers combined were estimated to be 14,000 for that year.11 The results presented in this analysis indicate that, throughout the last part of the last century, HCC mortality steadily increased. The discrepancy in previous estimates does point to the need for careful analysis of information collected on death certificates in determining mortality statistics.
In the present analysis, we found a 2.2-fold increase in the number of deaths and 67% increase in mortality rate between 1980 and 1998. The strength of this analysis in the assessment of mortality statistics is that the data used were relatively rich in detail, allowing maximum use of all information available. Thus, previous reports simply based on the "underlying cause of death" field of the Death Registry data either under- or overestimated the number of HCC deaths. The mortality rate calculated in this analysis closely track the incidence of HCC estimated using the SEER data. For example, the incidence rate for 1996-1998 by El-Serag et al3 was 3.0 per 100,000-year, compared with the mortality rate of 2.58 for 1998 in this study. Furthermore, the agreement in our estimates about in-hospital deaths between the Death Registry and NIS data lends strong support to their accuracy and validity.
This analysis was able to gain a partial insight into the underlying cause of increasing frequency of HCC. For example, in the NIS data, increasing proportions of HCC hospitalizations were associated with hepatitis B or C, which accounted for 28% of all HCC hospitalizations and almost 20% of all in-hospital deaths in 2000. A similar trend was noted in the mortality data, although ascertainment of underlying liver disease was much less complete. There is little doubt that part of this observed increase represents a detection bias in that testing for hepatitis C became available only in 1990, and many deaths related to hepatitis C had not been recognized as such before the early 1990s.12 Although the magnitude of this ascertainment problem is difficult to gauge, it appears likely that increasing numbers of people of long-standing hepatitis C infection is partly responsible for this trend.
In both Death Registry and NIS data, a large proportion of cases had no mention of other liver disease diagnosis. Undoubtedly, this observation is partly due to omission; however, it does raise a question whether there is an increasing frequency of HCC that is unrelated to obvious causes of liver disease such as viral hepatitis or alcoholic liver disease. One potential scenario may be nonalcoholic fatty liver disease, which may remain silent even when cirrhosis or HCC develops.13 There are many reports to indicate that a significant proportion of cryptogenic liver disease may indeed represent "burned out" NAFLD.14,15 There is widespread concern that the "epidemic" of obesity in this nation may lead to a large number of people with fatty liver disease.16 Perhaps, we might be already witnessing the early results of the NAFLD epidemic.
The racial distribution of HCC is also instructive. In terms of absolute numbers, the majority of HCC deaths were whites. However, the mortality rate was far higher in people of other races-almost 4 times higher than whites and 2 times higher than Africans. These HCC deaths are most likely due to long-standing hepatitis B infection, which underscores the importance of primary and secondary prevention efforts in this particular population segment.17 Although the implementation of universal vaccination for hepatitis B will essentially eliminate incident infections in native-born Americans,18 a recent influx of immigrants from hepatitis B endemic countries is likely to continue the increase in the number of patients with chronic hepatitis B infection virus developing HCC, and efforts at identification of individuals at risk and appropriate medical follow-up may be needed to reduce the future burden of HCC.18 Recent data from Taiwan indicate that antiviral therapy might be able to reduce the occurrence of HCC among those with advanced chronic hepatitis B.19
The differences in the temporal trend of HCC mortality across sex and racial categories shown in Table 1 are worth noting. In men, the trend in HCC mortality was similar across the 3 racial categories at 27% to 34% increase per decade. In women, the rise was slower (3%-20%) than in men, which indicates that the sex gap in HCC mortality is becoming even larger. This may be related to the fact that the prevalence of risk factors for HCC increased more sharply among men (eg, hepatitis C).20 More intriguingly, the rate of increase in HCC mortality among women was noticeably different across race. Although this may represent a spurious finding (eg, underreporting of HCC in minority women), it may also be a reflection of differences in risk factors (eg, increasing hepatitis C or NAFLD among white women). Further focused inquiry in this regard appears warranted.
The economic impact of HCC is considerable. A report by a task force commissioned by the American Gastroenterological Association (AGA) estimated the total cost associated with HCC to be $988 million in 1998, of which $562 million was attributed to hospital service utilization.21 Our analysis suggests that this was an overestimate because the estimated total charges in this analysis for 1998 were $272 million. This is probably related to the fact that the AGA analysis included all deaths coded to 155.2, thus including all primary and secondary liver cancers. It is important, however, to point out the progressive increase in hospital service utilization seen in this analysis. For example, there was a 40% increase in hospital charges between 1998 and 2000 alone.
The increases in in-hospital service utilization was in part due to employment of highly resource-intensive activities during the hospitalization. One such example is liver transplantation. In this analysis, there were 302 hospitalizations with liver transplantation in 2000. The United Network for Organ Sharing (UNOS) reports that 129 recipients underwent liver transplantation for HCC in the same year.22 We believe that the latter figure is an underestimate because, prior to the implementation of MELD-based allocation system, many cases with HCC were not reported as such to the UNOS.23 In addition to transplantation, local ablative therapy (such as percutaneous ethanol injection and radio-frequency ablation) was employed in 593 hospitalizations in 2000. Furthermore, the number of hospitalizations that involved angiography or embolization increased from 373 to 925 and hepatic resection from 379 to 824 between 1988 and 2000. Unfortunately, NIS does not allow longitudinal linkage of unique patients to evaluate patient survival in conjunction with these interventions. Although sporadic reports indicate prolongation of survival in patients undergoing some of these procedures,10 their true population-wide impact could not be estimated in this analysis.
There are several important limitations to this study. First, for both data sets that were used, we did not have the means to verify diagnoses accurately. The rigor with which death certificates are completed has been questioned, which may not be compensated for by post hoc data manipulation at the Centers for Disease Control and Prevention.24 Likewise, NIS is an administrative database, and diagnostic information is not verifiable (eg, by histology reports), although it appeared to be more complete than the Death Registry data. These limitations may result in underestimation (eg, HCC diagnosis not listed in patients who die of end-stage liver disease) or overestimation (eg, metastatic liver lesion presumed to be primary) of true HCC mortality. The same caveat applies to other diagnoses of liver disease that may have coexisted in the decedent. However, it may be worth pointing out that, regardless of the direction of any biases that these data may entail, the temporal trend is much more robust because those biases are less likely to shift over time.
With regard to the NIS data, the largest limitation is that the database records individual hospitalizations as opposed to individual patients, and it does not allow cross-linking of records by individual identifier, as a precaution to preserve patient confidentiality. Thus, we were not able to follow individual patients longitudinally or provide cumulative information on a given patient. We minimized the risk of introducing biases by restricting our analysis to 1-time events such as death or individuals' hospitalization charges. Finally, although we believe that the NIS data that we used provided a representative cross-sectional sample of HCC in the United States, health care in other sectors such as the Veterans Administration system was not included in the study.
These caveats not withstanding, relevant conclusions may be drawn from this study. First, during the relatively short time span in the last part of the 20th Century, the burden of HCC increased substantially in terms of mortality and hospital service utilization. Second, viral hepatitis seems to be an important reason for this increase, although hepatitis alone may not explain the increase. Third, significant differences were found in race-specific HCC mortality rates: Mortality rate was several-fold higher in nonwhite, non-African race. These data indicate that a number of convergent epidemiologic factors are contributing to the rising burden of HCC. To the degree that the prevalence of these conditions (chronic liver disease because of hepatitis B and C and possibly NAFLD) is expected to persist, if not increase, the burden of HCC is likely to continue in the future. Closer epidemiologic surveillance is needed to understand better the causation of HCC at the population level, and judicious use of that knowledge to implement primary and secondary prevention measures is urgently warranted.
|
|
| |
| |
|
|
|
