To determine mortality rates and risk factors among males with hemophilia who receive care in federally supported HTCs, we analyzed mortality data on the 85% of HTC patients voluntarily enrolled in UDC. Demographic, diagnostic, clinical, and treatment data were collected for all UDC patients on a standard form at annual comprehensive care visits. Proportion of expected annual UDC visits for each patient was the number of visits attended divided by the number of years between UDC enrollment and the end of the study. Blood specimens were tested centrally for markers of hepatitis and HIV infection. Data on mortality including date and causes of death was submitted by HTC staff using a standard form. Follow-up began at initial UDC visit and continued until date of death or December 31, 2006, whichever came first. Mortality rates were calculated as the number of deaths divided by person time and multiplied by 10,000 (deaths /104 person years PY). Univariate associations used the relative risk and multivariate analysis used Cox proportional hazards regression. Over the study period, 12,883 males with hemophilia ≥2 years old were enrolled in UDC and followed for a total of 68,060 PY. Among UDC participants there were 451 deaths for an overall mortality rate of 66.3 /104 PY. Hemophilia (14.2%), HIV (17.5%), and liver (29.1%) related causes of death were most commonly specified. As expected, mortality was strongly associated with age and ranged from 2.5 /104 PY for 2–10 year olds to 270 /104 PY for those >60 years. Also significantly associated with higher mortality in univariate analyses were Native American (NA) race vs. white, hemophilia A vs. B, severe vs. non-severe disease, HIV infection, hepatitis B active and past infection, hepatitis C infection, any restriction in activity level vs. none, increasing body mass index (BMI), history of intracranial hemorrhage (ICH), alcohol abuse, and any sign of liver disease including jaundice, ascites, varices, or elevated ALT/AST or PT. On the other hand, patients who attended from 1/3 to 2/3 of their expected UDC visits were 60% less likely to die and those who attended more than 2/3 of expected UDC visits were 90% less likely to die than those patients who attended less than 1/3 of their expected UDC visits. The distribution of patients attending <1/3, 1/3 to 2/3 and >2/3 of expected visits was 15.3%, 31.3% and 53.4%, respectively. A multivariate model including all risk factors indicated that NA race RR=2.6 (95% CI: 1.2, 5.6), HIV infection RR=3.3 (2.6, 4.3), severe disease RR=1.8 (1.4, 2.3), alcohol abuse RR=1.3 (1.0, 1.7), jaundice RR=2.0 (1.4, 3.0), ascites RR=1.5 (1.0, 2.4), varices RR=2.0 (1.3, 3.1), elevated ALT/AST RR=1.4 (1.1, 1.8), and elevated PT RR=1.5 (1.1, 2.1) increased risk for mortality. Whereas attending 1/3 to 2/3 expected visits RR=0.36 (0.29, 0.44), attending >2/3 expected visits RR=0.07 (0.05, 0.10) and 10 unit increase in BMI RR=0.72 (0.58, 0.90) were protective against mortality. Hemophilia type, activity restriction, history of ICH, and hepatitis B and C infection were no longer significantly associated with death in the multivariate model. Our finding that patients enrolled in the UDC project who returned regularly to HTCs for comprehensive care were much less likely to die, even after adjusting for other mortality risk factors, supports results of previous studies that showed lower mortality among patients receiving comprehensive care in HTCs.
Disclosure: No relevant conflicts of interest to declare.