Background: Hereditary hemorrhagic telangiectasia (HHT) is the second most common hereditary bleeding disorder worldwide. Patients with HHT develop vascular anomalies in multiple organs that cause complications such as bleeding, thrombosis, anemia and high-output heart failure, resulting in poor quality of life. Few studies define the prevalence, risk factors, or screening tools for bleeding in HHT.

Methods: We conducted a cross-sectional study of HHT patient discharges in the National Inpatient Sample (NIS) database from 2016-2018. Types of bleeding and other medical conditions were identified using ICD-10 codes. Analysis included NIS-provided discharge-level weights to reflect national estimates. NIS is the largest publicly available inpatient database in the U.S. developed by Healthcare Cost and Utilization Project (HCUP), containing a 20% stratified sample of U.S. hospital discharges. Categorical and continuous variables were assessed for association with HHT and bleeding status by Rao-Scott Chi-Square and weighted simple linear regression, respectively. Weighted multivariable logistic regression was performed to determine bleeding risk factors in HHT patients. Variables included gender, race, hypertension (HTN), chronic obstructive pulmonary disease (COPD), liver disease and cirrhosis, peptic ulcer disease (PUD), portal HTN, peripheral vascular disease (PVD), malignancy, hepatitis C (HCV), portal vein thrombosis (PVT) and Charlson Comorbidity Index (CCI).

Results: Among 21,393,912 adult discharges, a total of 2591 discharges with HHT (0.012%) were identified. Compared with HHT without bleeding (HHT-NB), HHT discharges with bleeding (HHT-B) were more likely to be African American (15.8% vs 12.6%, p=0.046), male (44.4% vs 39.4%, p=0.03), have HTN (56.9% vs 51.0%, p=0.01); liver disease (21.6% vs 12%), cirrhosis (9.8% vs 5.4%), PUD (8.7% vs 1.0%), portal HTN (7.1% vs 3.0%), HCV (3.3% vs 1.3%), PVT (1.5% vs 0.3%), each p<0.001; PVD (6.9% vs 4.8%, p=0.03), and higher morbidity (CCI) (mean ± sd: 2.3 ± 0.09 vs 2.1 ± 0.05, p=0.03), while malignancy (3.9% vs 6.8%, p=0.008) was less common. Additionally, congestive heart failure (CHF) (26.1% vs 18.0%), gastroesophageal reflux disease (GERD) (25.5% vs 19.8%), liver disease (21.6% vs 12.3%), cirrhosis (9.8% vs 3.5%), portal HTN (7.1% vs 3.9%) and HCV (3.3% vs 1.3%), each p<0.001, were each more common in HHT-B than in non-HHT with bleeding (non-HHT-B). GI bleeding in HHT was as likely to occur in the upper GI tract as in non-HHT discharges (35.3% vs 36.7%, p=0.47), but significantly less likely in the lower GI tract (19.5% vs 29.8%, p<0.001). The most common type of bleeding in HHT were AV malformation (AVM) (32.1% vs 1.3%), epistaxis (17.7% vs 0.6%), and telangiectasia (2.3% vs 0.2%), each p<0.001, compared to non-HHT. Surgical/procedural bleeding (0.8% vs 4.2%), postpartum hemorrhage (PPH) (0.3% vs 6.5%), heavy menstrual bleed (HMB) (0.9% vs 5.9%), and malignancy (3.9% vs 10%) were less common in HHT vs non-HHT, each p<0.001, but iron deficiency anemia (IDA) was more common, 15.5% vs 7.5%, p<0.001, despite lower use of aspirin (4.7% vs 13.2%) or anticoagulants (5.1% vs 10.5%), each p<0.001, while mean length of stay (5.6 days vs 6.1 days, p=0.04) and inpatient mortality (1.8% vs 3.6%, p=0.01) were lower in the HHT group. In multivariable logistic regression, correlates of bleeding in HHT included PUD, adjusted odds ratio (AOR) 8.77 [95%CI 5.07 - 15.2]; PVT, AOR 3.81 [1.49 - 9.79]; and HCV, AOR 2.19 [1.17 - 4.08], while COPD, AOR 0.77 [0.61 - 0.98]; and malignancy, AOR 0.58 [0.36 - 0.94] were protective. Factors most commonly associated with HHT-B included CHF, AOR 2.57 [2.00 - 3.30]; and IDA, AOR 2.24 [1.82 - 2.77] compared to non-HHT-B.

Conclusion: Among those with HHT, those with bleeding were more likely to be male, African American, and have HTN, PUD, and HCV than those without bleeding. GI tract AVM and angiodysplasia were the most common cause of bleeding in HHT. IDA was more common in HHT, but surgical and gynecological bleeding such as PPH and HMB were significantly less common compared to non-HHT. The presence of GI bleeding associated with AVM or angiodysplasia should raise the possibility of HHT. IDA was equally common in HHT with and without bleeding (15.5%) suggesting either lack of testing, poor response to therapy, or ongoing losses with unrecognized bleeding.

Ragni:Takeda: Honoraria, Other: Provided study (rVWF) drug for trial; BioMarin: Honoraria.

Author notes

This icon denotes a clinically relevant abstract

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution