Burden of Illness Among Pediatric Hemophilia Patients with and without Central Venous Access Devices Treated in U.S. Hospitals

Introduction:  Central venous access devices (CVADs) are valuable tools to facilitate repeated or urgent treatments for hemophilia patients. However, CVADs may increase the risk for local and systemic infection, as well as thrombosis. There is limited real-world information on the use of CVADs among the pediatric hemophilia population treated in a hospital setting. The aim of this study was to describe and compare the characteristics and burden of illness of the hemophilia population with and without CVADs.

Methods: This observational study used data from the Premier research database and included inpatients and outpatients who were age ≤ 18 years at the time of index admission and discharged during 2006-2014 from one of 548 U.S. hospitals. We identified hemophilia patients using principal or secondary ICD-9 diagnosis codes 286.0x or 286.1x, and CVAD exposure using billing information that represented the procedure with or without a corresponding ICD-9 procedure code of 86.07. We matched CVAD to non-CVAD exposed patients using a 1:1 matching scheme based on age (+/- 1 year), Factor-VIII therapy (yes/no), Factor-IX therapy (yes/no), number of prior HCP encounters in previous 365 days (+/- 5), previous hospitalization for infection in previous 365 days (yes/no), and geography (9 census region categories). Comparisons of outcomes including infections, thrombosis, revisit rates, and length of stay between groups were made using Wilcoxon signed rank tests for medians and generalized linear models for categorical variables to account for correlation between matched pairs, stratified by inpatients and outpatients.

Results: There were a total of 4,793 eligible pediatric hemophilia patients during the study period. Of these, 197 (4.1%) had CVAD exposure according to our criteria. The matched sample included 163 CVAD cases and 163 controls (142 inpatients and 184 outpatients).

The inpatient hemophilia population was median age 5 years, primarily male (93%),  white (55%), treated in urban hospitals (94%) in the south region (46%), and had Medicaid as the primary payer (47%). Compared with the non-CVAD group, patients with CVAD exposure in the pre-matched group were more often male (98% vs. 92%), younger (3 vs. 5 years), and treated in urban (98% vs. 94%), teaching hospitals (79% vs. 64%). Among the inpatient matched sample, CVAD cases compared with controls experienced a greater frequency of all-cause 30-day (24% vs. 11%, p=0.03), 60-day (31% % vs. 13%, p=0.01), and 90-day (33% vs. 17%, p=0.02) outpatient revisits, and more all-cause infections (12% vs. 2%, p=0.02), but no notable difference in the frequency of thrombosis (4% vs. 2%, p=0.48) or inpatient revisits. The median hospital length of stay for inpatients was greater for CVAD cases compared with controls (4 vs. 3 days, p=0.01).

The outpatient hemophilia population was slightly older (median age 9 years) and less often male (82%) or treated in urban hospitals (82%), but otherwise had similar demographic characteristics to the inpatient population; 58% white, with 49% in the South and 45% having Medicaid as the primary payer.  Compared with the non-CVAD group, outpatients with CVAD exposure in the pre-matched group were more often male (97% vs. 82%), younger (4 vs. 9 years), and treated in urban (92% vs 82%), teaching hospitals (61% vs. 39%). Among the outpatient matched sample, CVAD cases compared with controls had greater frequencies of subsequent inpatient and outpatient visits at 30-, 60-, and 90-days (all p< 0.05), as well as all-cause infections (24% vs. 3%, p=0.0001) and thrombosis (9% vs. 0.9%, p=0.03).

Conclusions: CVAD exposure was associated with a greater frequency of infections, thrombosis, and inpatient and outpatient revisit rates, as well as longer length of stay, among pediatric hemophilia patients treated in U.S. hospitals. The results of this study may inform research efforts for novel treatments best suited to the unique needs of hemophilia patients with CVADs.