Abstract

Polymerization of hemoglobin in sickle cell anemia (HbSS) results in red blood cell damage, vasoocclusive disease and hemolytic anemia. Plasma hemoglobin and arginase released during hemolysis produce endothelial dysfunction and lead to vascular complications such as leg ulceration, priapism, pulmonary hypertension and increased risk of death. Examination of 2 independent patient groups from the NIH and CSSCD revealed considerable heterogeneity in the rates of hemolysis as estimated by LDH. We therefore hypothesized that inter-individual variation in hemolytic rate influences the phenotypic expression of HbSS. To test this hypothesis, we have defined a hemolysis phenotype as a clinical variable in our 2 HbSS groups. At NIH, LDH was analyzed as a quantitative trait in 166 adults to define a subgroup above the 75th percentile suggestive of exaggerated hemolysis (n= 42, mean LDH 645). When compared with 42 cases in the lowest LDH quartile (mean 230), the hyperhemolysis group had significantly lower hemoglobin and higher AST, bilirubin, and absolute reticulocyte levels supporting this phenotypic classification. HbF, arginine:ornithine ratios and arginase 1 activity were also significantly different. Clinically, the hyperhemolysis group had an increased prevalence of leg ulcers (OR 5.10; P=0.007), priapism (OR 3.67; P=0.16), and pulmonary hypertension (TRV > 3.0 m/s; OR 12.00; P<0.001). In addition, these patients have fewer severe vasoocclusive pain crises (3.1 ER visits/yr. with high LDH vs. 9.1/yr. with low LDH; P=0.004), supported by a lower prescription rate for hydroxyurea (37% with high LDH vs. 61% with low LDH; P=0.06). To validate these findings, we repeated the LDH phenotypic assignment analysis using 451 comparable CSSCD adults greater than 30 years of age. An identical pattern of laboratory results was seen when comparing high and low LDH quartiles. Hyperhemolysis was defined by anemia, elevated total bilirubin and AST and lower HbF. The most striking observations for this group were the significantly higher proportion of males (55% vs. 27% with low LDH; P<0.001), occurrence of leg ulcers (OR 3.44; P<0.001), and frequent death (33% vs. 16% with low LDH; P=0.003). There were also trends towards more prevalent priapism and stroke in the hemolysis group. After adjustment for HbF and gender, the associations between hyperhemolysis and risk of leg ulceration and risk of death remained significant in the CSSCD. Together, these findings suggest that excessive hemolysis results in a vascular HbSS phenotype characterized by leg ulcers, priapism, pulmonary hypertension, possibly stroke, and death. The lack of an association between excess hemolysis and pain/ACS suggests this phenotype is distinct from vasoocclusion/hyperviscosity. The consistent association of hyperhemolysis with this spectrum of symptoms, even after HbF correction, indicates the existence of unknown determinants for hemolysis. Future environmental and genetic studies of hyperhemolysis may provide novel mechanistic insights into the pathogenesis and treatment of HbSS vascular complications. From a clinical standpoint, these data further support the use of LDH as a simple biomarker for identifying a sub-set of HbSS patients at high risk for hemolysis-driven complications and death, who may otherwise be overlooked by clinicians because of infrequent vasoocclusive pain events.

Disclosure: No relevant conflicts of interest to declare.

Author notes

*

Corresponding author