Background/Purpose: To our knowledge, only a few population-based studies for vasculitis mortality exist; and most are limited to vasculitis sub-types. Therefore, our study objective was to assess time-trends in in-hospital mortality in vasculitis and compare it to the general population, using the U.S. National Inpatient Sample (NIS).

Methods: We included the data from the Healthcare Cost and Utilization Project (HCUP) NIS from 1998 to 2014. The NIS is a de-identified national all-payer inpatient health care database that has a 20% stratified sample of hospital discharges. NIS is used frequently for creating U.S. national estimates.

Cases of hospitalized vasculitis were defined based on the listing of 446.xx or 447.6 as the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnostic codes in the primary diagnosis position. A previous study showed a sensitivity of 93% and specificity of 95% using this approach. We limited our study to 2014, the last year ICD-9-CM codes were used in the U.S. before switching to ICD-10-CM in 2015.

We calculated the unadjusted in-hospital mortality rates per 1,000 hospitalizations for people with versus without primary vasculitis hospitalizations, with the respective denominators. We calculated age-adjusted rates by grouping age in quintiles (< 20, 20-39, 40-59, 60-79, >79) and age- and sex-adjusted rates. We analyzed in-hospital mortality rate time-trends using the Cochran Armitage test, weighted by the number of hospitalizations each year.

Results: There were 266,461 primary vasculitis hospitalizations in 1998-2014 with 7,215 in-hospital deaths (2.7%). Mean age was 42.9 years, 57% were men, 50% were white, 30% had a Deyo-Charlson comorbidity score of ≥2 and 18% had a Medicaid payer (Table 1).

Unadjusted in-hospital mortality in primary vasculitis hospitalizations decreased by 43% from 32.8 per 1,000 in 1998 to 18.7 per 1,000 in 2014 (p< 0.01), compared to a 24.5% reduction in deaths for all NIS claims without vasculitis (28.1 to 21.2 per 1,000; p< 0.01; Figure 1).

Age- and sex-adjusted in-hospital mortality decreased both in primary vasculitis hospitalizations from 27.3 per 1,000 claims in 1998 to 19.1 in 2014 and non-vasculitis hospitalizations from 15.1 to 13.2 per 1,000 in 2014, respectively (p< 0.01; Figure 1; Table 2). The age- and sex-adjusted mortality rate gap between those with versus without vasculitis narrowed (Figure 1); vasculitis to non-vasculitis in-hospital mortality ratio was 1.81 in 1998-2000 and 1.45 in 2013-2014 (Table 2). Primary vasculitis hospitalizations per year remained fairly constant from 1998-2014 at 14-18,000 per year; rate was 43 per 100,000 NIS claims in 1998-2000 versus 43.2 in 2013-14; i.e., 0.05% of claims per year.

Conclusion: Adjusted in-hospital mortality decreased significantly over time in both the general U.S. population and primary vasculitis hospitalization cohorts, and the absolute and relative reductions were larger for vasculitis hospitalizations. Age- and sex-adjusted in-hospital mortality was higher in vasculitis compared to the general population in both 1998 and 2014, but the gap narrowed over time.

Figure 1. Time-trends in unadjusted (A) and age- and sex-adjusted (B) in-hospital mortality rates per 1,000 population of hospitalizations in people with vasculitis compared to people without vasculitis. Figure 1 legend The y-axis shows rate per 1,000 hospitalization claims with each respective denominator

Table 1. Demographic and clinical characteristics of people hospitalized with vasculitis in the U.S.

Table 2. Mortality rates per 1,000 population in people hospitalized without versus with Vasculitis: unadjusted based on the NIS sample; age- and age- and sex-adjusted rates based on the NIS sample and the census data

« Back to ACR Convergence 2020

ACR Meeting Abstracts - https://acrabstracts.org/abstract/declining-in-hospital-mortality-in-vasculitis-a-17-year-u-s-national-study/