During 2000–2019, there were an estimated 1 286 397 HCV-related hospitalizations in the US (Table 1). The annual estimated number of HCV-related hospitalizations was lowest in 2000 (41 232) and highest in 2012 (79 150), while the annual age-adjusted HCV-related hospitalization rate was lowest in 2019 (18.7 per 100 000 population) and highest in 2012 (29.6 per 100 000 population) (Figure 1). Most HCV-related hospitalizations occurred among persons aged 45–64 years (71.8%), males (67.1%), White non-Hispanic persons (60.5%), urban residents (85.0%), persons with Medicaid or Medicare (64.0%), and persons with median household income ≤50th percentile (61.3%) (Table 1).

Estimated age-adjusted rates of HCV-related hospitalizations by demographic characteristics appear in Figure 2. The rate was highest among persons aged 45–54 years during 2000–2013 and then among persons aged 55–64 years from 2014 to 2019. In terms of insurance type, the rate was highest for Medicaid recipients during all but 2 years of the analysis period (2000, when private insurance recipients had the highest rate, and 2013, when Medicare recipients had the highest rate). The rate was highest among White non-Hispanic persons, followed by Hispanic persons and then Black non-Hispanic persons throughout the analytic period. The rate was higher among males than females throughout the analytic period.

Two joinpoints (in 2003 and 2013) and 3 segments were identified in the joinpoint regression analysis (Figure 3). The national age-adjusted HCV-related hospitalization rate increased significantly during 2000–2003 (APC, 9.4% [95% CI, 5.5%–13.5%]; P < .001) and 2003–2013 (APC, 1.8% [95% CI, 1.3%–2.4%]; P < .001) before decreasing significantly during 2013–2019 (APC, −7.6% [95% CI, −8.5% to −6.6%]; P < .001) (Figure 3).

Percentage change in the annual age-adjusted HCV-related hospitalization rates by selected demographic characteristics appear in Table 2. Comparing 2000 to 2019, persons aged 55–64 years experienced a 132.9% increase in HCV-related hospitalization rate, while persons aged 35–44 years and 45–54 years experienced rate decreases of −47.0% and −43.7%, respectively. From 2000 to 2019, the highest increase in HCV-related hospitalization rate occurred among Black non-Hispanic persons (22.3%), followed by White non-Hispanic persons (18.1%); the largest decline in hospitalization rate occurred among Asian/Pacific Islander persons (−22.5%). Comparing 2000 to 2019, HCV-related hospitalization rates increased for Medicaid (41.6%) and Medicare (9.5%), while rates for private insurance and uninsured declined −57.1% and −10.7%, respectively. Age-adjusted HCV-related hospitalization rates declined for every demographic category examined when comparing rates in 2013 to rates in 2019.

In this analysis of 20 years of pre-COVID-19 pandemic NIS data, we found that HCV-related hospitalizations predominantly occurred among persons aged 45–64 years, males, White non-Hispanic persons, and persons with Medicaid or Medicare insurance. Overall HCV-related hospitalization rates increased significantly prior to the availability of DAAs during 2000–2013, and then decreased significantly during 2013–2019 after DAAs became available; the 2019 age-adjusted HCV-related hospitalization rate was the lowest in 20 years. However, the decline in hospitalization rate was not observed equally among all age, race/ethnicity, or insurance categories.

We observed a decline in the trajectory of the increase (ie, a reduction in slope) in the age-adjusted HCV-related hospitalization rates beginning in 2003. The introduction of pegylated interferon as a hepatitis C treatment in 2001 might have contributed to the change in rates observed in 2003. Prior studies have demonstrated associations between achieving SVR from interferon-based treatments and hepatic benefits, including hepatic fibrosis regression, cirrhosis resolution, decreased portal hypertension complications, and lower liver failure risk [21-24]. Additionally, several recent studies support the overall trend of decreased HCV-related hospitalizations in the era of DAAs observed in our analysis. In Canada, researchers analyzed administrative and pharmacy data and found a modest decline in HCV- and liver-related hospitalizations following significant uptake of second-generation DAAs [25]. In the US, the Chronic Hepatitis Cohort Study (CHeCS) demonstrated that patients with chronic hepatitis C who achieved SVR after DAA treatment had significantly lower rates of all-cause and liver-related hospitalizations compared with chronic hepatitis C patients who had not received DAA treatment [26]. Two large cohort studies from the Veterans Administration published in 2017 demonstrated that achieving SVR from DAA treatment was associated with a reduced risk of hepatocellular carcinoma [27, 28]. Additionally, a cohort study of 245 596 adults with chronic hepatitis C found that DAA treatment was independently associated with a lower risk of mortality and liver- and non-liver-related outcomes compared with no treatment [29].

Prior to 2013, the age-adjusted HCV-related hospitalization rate was highest among persons aged 45–54 years. However, persons aged 55–64 years experienced the largest increase in HCV-related hospitalization rate during the pre-DAA portion of the analysis period and ended up as the age group with the highest age-adjusted HCV-related hospitalization rate after 2013. Although the HCV-related hospitalization rates declined for all age groups during 2013–2019, the declines were notably smaller among those 35–44 years old and 18–34 years old during this period, which is likely a reflection of increasing HCV infection incidence and prevalence in those age groups [4]. Improving access to curative DAA agents among younger persons more recently infected has the potential to interrupt further HCV transmission and prevent future morbidity that could develop if recent HCV infections are left untreated.

Age-adjusted HCV-related hospitalization rates were highest among White non-Hispanic persons throughout the entire analytic period and were approximately 3-fold higher than the rates among Hispanic persons and 4-fold higher than the rates among Black non-Hispanic persons. Despite these comparatively lower rates, however, the rate of HCV-related hospitalizations among Black non-Hispanic persons increased 22.3% during 2000 to 2019, which was the largest increase among all race/ethnicity categories. An explanation for this shift could be reduced or delayed access to DAAs among Black non-Hispanic persons relative to other racial/ethnic categories. This is to some degree supported by cohort analyses conducted after DAAs were released that indicated racial differences in frequencies of initiation. Notably, a cohort study of hepatitis C patients treated at the Veterans Administration found that Black patients had lower odds of receiving treatment in the DAA era than White patients, while a CHeCS analysis found that Black non-Hispanic race/ethnicity was associated with noninitiation of DAA treatment [30, 31]. Disparities in access to treatment could have led to an accumulation of untreated HCV-infected Black non-Hispanic persons in the US who progressed to severe liver disease, requiring HCV-related hospitalization, compared with persons who had earlier access to DAA treatment. Further research is warranted to identify the causes of these race/ethnicity differences, including the role of differential access to healthcare in general and DAAs in particular.

The temporal shift to Medicaid and Medicare for insurance coverage of HCV-related hospitalizations is a noteworthy finding that can influence future health policy decisions. During the analysis period, the rates of HCV-related hospitalizations paid for by private insurance decreased, while the rates of hospitalizations paid for by Medicaid or Medicare increased. In contrast to expenditures for hospitalization, a recent analysis of national pharmacy claims data demonstrated that during 2014–2020, Medicare and Medicaid paid the pharmacy claims for 54% of persons treated with DAAs [32]. In 2014, the Affordable Care Act (ACA) was implemented in the US, expanding health insurance coverage and reducing uninsured rates [33]. Implementation of the ACA is likely partially responsible for the uptick in the rates of HCV-related hospitalizations paid for by Medicaid from 2013 to 2014, and the concomitant decline in the rates of hospitalizations among uninsured persons. Although many US states have expanded DAA access for Medicaid recipients, restrictions related to prescriber type, substance use, and hepatic fibrosis persist [34]. Given the well-documented benefits of treating HCV infection early to prevent end-stage liver disease [6], optimization of DAA treatment expenditures by public and private payers is likely to reduce the future burden of disease and related medical costs.

Alternative classification schemes might have produced different estimates of the HCV-related hospitalization burden. We performed a sensitivity analysis (data not shown) examining the annual proportion of overall adult hospitalizations attributable to several different HCV-related classification schemes. During the analysis period, the proportion of adult hospitalizations with hepatitis C as a primary discharge diagnosis ranged from a peak of 0.11% in 2011–2013 to a low of 0.01% in 2017–2019; the proportion with hepatitis C classified according to the definition used in our analysis ranged from a peak of 0.26% in 2012–2014 to a low of 0.14% in 2000; and the proportion with hepatitis C as any discharge diagnosis ranged from a peak of 2.19% in 2015 to a low of 0.80% in 2000. Our goal was to produce a conservative estimate of the hospitalization burden that could legitimately be attributed to hepatitis C; utilizing our classification scheme, which was geared toward capturing end-stage manifestations of HCV infection (eg, more sensitive than solely using hospitalizations with hepatitis C as a primary discharge diagnosis and more specific than using hospitalizations with hepatitis C as any discharge diagnosis) accomplished this aim. Additionally, we justified inclusion of alcoholic liver disease as a liver disease–related condition in our HCV-related hospitalization definition due to the pervasiveness of alcohol use among patients with hepatitis C and the inability to rule out contributions of HCV-mediated liver injury to hospitalizations where alcoholic liver disease was the primary diagnosis and hepatitis C was a secondary diagnosis.

Our findings are subject to limitations. First, the unit of observation in the NIS is a hospitalization, not a patient; readmissions for the same patient are counted as new hospitalizations. Persons with advanced liver disease typically require intensive medical management; hospital readmissions during a single calendar year would not be uncommon [8, 35-39]. If a small number of patients accounted for a large number of hospitalizations, they could have disproportionately affected the distribution of demographic characteristics in our analysis, or potentially contributed to an overestimation of the HCV-related hospitalization rates. Although there was a sustained decline in both the estimated number and age-adjusted rates of HCV-related hospitalizations following the introduction of DAAs, other factors, including HCV-related mortality, might have contributed to the decline in HCV-related hospitalizations after 2013. In the context of this analysis, deaths occurring among patients with advanced liver disease would reduce in number the hospitalization-prone segment of the prevalent population and consequently lower the HCV-related hospitalization rate; the magnitude of such an effect would depend on how frequently such a decedent was readmitted in a calendar year (data not available in NIS). Second, changes in coding practices during the analysis period might have influenced the estimated number and rates of HCV-related hospitalizations observed. However, we would expect that any such influence would have been in the direction of increased HCV-related hospitalizations due to increased awareness of hepatitis C over time. Third, this was a retrospective, cross-sectional analysis; causality cannot be inferred. Fourth, despite the NIS being the largest publicly available all-payer inpatient database, our analysis likely underestimates the HCV-related hospitalization burden in the US due to the omission within the NIS of hospitalization data from the Veterans Administration. Finally, we did not include extrahepatic manifestations of HCV infection in the liver disease–related conditions we assessed as part of our HCV-related hospitalization definition. However, a prior NIS study found that the prevalence of extrahepatic manifestations among hospitalizations involving a hepatitis C diagnosis code was <0.5% during 2004–2011, suggesting that including extrahepatic manifestations would not have substantively altered the results of our analysis [40].

Although multiple factors likely contributed, overall HCV-related hospitalization rates declined steadily after 2013, coinciding with the release of DAAs. However, the declines were not observed equally among age, race/ethnicity, or insurance categories. Expanded access to DAA treatment is needed, particularly among Medicaid and Medicare recipients, to reduce morbidity and mortality, prevent further transmission, eliminate disparities, decrease the healthcare and financial burdens associated with HCV-related hospitalizations, and ultimately eliminate hepatitis C as a public health threat.