Steady State Predictors of Mortality in Adults with Sickle Cell Disease

Background: With improved treatment modalities, longevity for sickle cell disease (SCD) is increasing. Yet there is still a subgroup of adult patients whose survival rate has not improved. These patients, if well defined by evidence based decisions, might be those for whom more aggressive disease modifying therapy would be appropriate.

Methods: We identified all patients with SS/Sβ⁰ seen at our medical center in 2002 whom we were still following in 2012 as well as those who had died between 2002 and 2012. Patients with SC and Sβ⁺ thalassemia were excluded. ED, clinic, and inpatient admissions over the entire years 2002 and 2012 were obtained. All 2002 and 2012 steady state parameters for a given laboratory test for a given patient were averaged. “Steady State” was defined as those values not within one day of an ED visit or within one week of a hospital admission. Data were assessed for normality; parametric and non-parametric bivariate tests were performed as appropriate. Data from 2002 and 2012 was compared with paired-T tests. Mortality data were analyzed using Kaplan-Meier curves and Cox proportional hazard models.

Results: We identified 289 SS/Sβ⁰ patients in 2012 (ages 18-87, 54% female) who had been present in our system in 2002 (survival cohort). We also identified 70 patients present in 2002 (ages 19-88 at death, 47% female) who had died between 2002 and 2012, inclusive (mortality cohort). Average age at death was 42.1±14.0 years, median survival was 58.3 years (95% CI: 54.5 – 63.0). Survivors had, in 2002, higher HbF, Hb, higher MCHC, lower white blood cell (WBC) counts and creatinine (Cr) levels, and fewer admissions than those who did not survive past 2012 (Table 1). Absolute Reticulocyte count (Retic), MCV, LDH, liver function tests, ED and clinic use were not significantly different between survivor and mortality groups. Cox proportional hazards model showed increased hazard ratios with lower HbF, higher WBC, Cr and increased admissions. When the survivors from 2002 were compared to themselves in 2012, they were shown to decrease HbF, Hb, MCHC, WBC, alkphos, and total bilirubin over time and to increase admissions, ED visits, Retic, MCV, Cr, and weight. Clinic visits, direct bilirubin and LDH were not significantly different.

Conclusions: As therapeutic alternatives increase, it may be important to tailor therapy, and the need for better prognostic markers becomes ever more important. We observed here that several known mortality predictors evolve over time. As the “survivor” population ages, they may begin to resemble those with a poorer prognosis with decreased Hb and HbF levels but increased admissions and ED visits. However in our sample, WBC, a recognized factor in mortality and morbidity, decreased further over time in those who survived. Similarly, absolute reticulocyte count increased further over time in the survivors, suggesting an ability of the marrow to respond to the increased anemia. Further studies should be done to distinguish which biomarkers, at what level and in what age groups, are truly predictive of severity in sickle cell disease.

*=p<.05, #=p<.01