National Hopitalization Rates for Sickle Cell Disease in the Hydroxyurea Era

Sickle cell disease (SCD) is a severe inherited hematologic disorder caused by the substitution of valine for glutamic acid on the beta-globin chain of hemoglobin, resulting in sickle hemoglobin (HbS). Under deoxygenation, HbS polymerizes into stiff fibers that cause red blood cell membrane changes, vaso-occlusion and numerous deleterious downstream events. Although the polymerization of HbS remains the principal event in the molecular pathogenesis of SCD, additional cellular events including endothelial cell and platelet activation and heightened leukocyte adhesion are all likely to contribute to in vivo vaso-occlusion. The primary effect of HbF in inhibiting HbS polymerization is through the formation of asymmetrical hybrids α₂β^Sγ which participate less readily in polymer formation. The beneficial effect of hydroxyurea (HU) in SCD is primarily due to induction of Hb F, thereby increasing length of time to polymerization. The relative importance of the additional effects in the efficacy of HU is unknown. In the Multicenter Study of Hydroxyurea, a Phase III multi-center randomized placebo-controlled trial of hydroxyurea in patients with SCD, HU showed a reduction hospitalizations for SCD, and an overall reduction in morbidity and was approved for therapy in SCD in adults in 1998. HU remains the only drug with FDA approval for prevention of sickle vaso-occlusive crises. We assessed the impact of HU on SCD, we looked at the trend in hospitalization rates in the US since FDA approval of HU for SCD in adults. We used the Nationwide Inpatient Sample (NIS), a stratified random sample of acute care hospitals in the United States containing about 8 million hospital stays each year, weighted to be representative of the US population, years 1998 to 2004. We identified discharges associated with a diagnosis of SCD by ICD-9-CM codes and limited our analysis to blacks alone. We looked at rates of hospitalizations incorporating population data from the US Census Bureau for each year 1998 – 2004. Poisson test for trend was used to test the trends in hospitalization rates over the years of study and SAS 9.1 for analysis. Between 1998 and 2004 the NIS included data representing between 92846 at the lowest and 119297 at the highest hospital stays in the US for each year. Approximately 70% of hospitalizations were primarily for a diagnosis of SCD. Ninety percent of discharges were in blacks, 3% in Hispanics and 2% in others. We looked at hospitalizations in blacks with SCD where there were 73,089 discharges associated with SCD in 1998 and 82,313 discharges associated with SCD in 2004. The corresponding rates of hospitalization of blacks for SCD in 1998 and 2004 were 223 and 229 per 100,000 blacks respectively (p<0.01) and using Poisson regression to test for a trend in hospitalization rates from 1998 to 2004, the p value was <0.01. The rates of the intervening years ranged from 197 to 233 per 100,000. From the data we conclude that on a national level, though the p values show statistical significance, the raw data shows clinically insignificant changes in hospitalization rates over the years studied, and if anything, a relative rise in hospitalization rates since 2001. A similar but smaller study looking at hospitalization rates in the state of Maryland documented an increase in hospitalizations between 1995 and 2003 and that 70% of sickle cell patients who were appropriate candidates for HU were not on the medication. With the scientifically sound pre-clinical data on HU in SCD, is it likely that this indicates that HU use has not been maximized. Guidelines on the indications, dosing, contra-indications and complications for specific age groups need to be developed and presented as a comprehensive document. In the recently dissolved structure of sickle cell care, the Comprehensive sickle cell centers cared for only 12% of the 80,000 patients with SCD in the US. By default a significant number of patients with SCD were and are cared for mostly in Emergency rooms and acute care settings which provide no opportunity for HU institution. This, as well as smaller studies, highlight the difficulty in replicating findings in clinical research to the real world. Further studies are needed to investigate the barriers to HU use by doctors and patients. This is particularly important as the NHLBI realigns its SCD Research Program, so that we benefit maximally in clinical practice from the data that we have already have and will generate from the new SCD program.