Hydroxyurea, the 1st line treatment for SCD, induces HbF, inhibiting polymerization of mutant HbS and cell sickling. However, laboratory tests that assess HbF induction do not quantify the % of effectively protected RBCs (%pRBC), i.e. those with HbF content (cHbF)>10 pg/RBC. While chromatography (HPLC) determines the HbF% among blood Hb (HbF%), classical flow cytometry (cFC) provides the % HbF+ RBCs (Fcell%), including RBCs with very little HbF. We have developed an unprecedented method of quantitative CF (qFC) that uses an internal cellular control, composed of samples of umbilical cord blood (UCB) and adult blood with known mean HbF contents (MCHbF), allowing the distribution of cHbF/RBC in the population is determined, including the %pRBC.

Validate the method and demonstrate its potential clinical application in SCD.

The MCHbF of 9 UCB samples were pre-calculated, multiplying the HbF% by the respective MCH obtained with a hematological analyzer, and compared with the MCHbF derived by qFC, in order to evaluate the correlation and linearity of the method. A total of 40 samples from 29 SCD (SS) patients had their hematimetric indices and %pRBC determined. From these, %Fcell (n = 27) and HbF% (n = 15) were obtained, totaling 12 samples characterized by all methods, which were evaluated for correlation between themselves and with hematimetric data. Hb<9/dL was used to distinguish high (n = 18) from low (n = 9) clinical risk patients. %Fcell and %pRBC were compared for their ability to predict high-risk patients, based on the area under the ROC curve (AUC).

Pre-calculated and qFC-derived MCHbF from UCB samples were significantly correlated (Pearson r = 0.65, p = 0.02). This correlation improved for fresh samples (n = 5; r = 0.91, p = 0.01), with pre-calculated and qFC-derived MCHbF differing by less than 10%. Of the 12 samples characterized by all methods, all 10 with HbF%>10% had %Fcell>99%. In contrast, %pRBC correlated strongly and linearly across all HbF% (r = 0.89, p < 0.0001). In line, both HbF% and %pRBC showed significant positive correlations with RBC counts, Hb, and hematocrit (Hct); and negative correlations with the % reticulocytes (%Ret) and Red Cell Distribution Width (RDW). However, %Fcell correlated only with Hct. Considering all 27 samples evaluated, %Fcell correlated further only with Hb and RDW; however, %pRBC was further correlated positively with MCHC, and negatively with WB cells, Ret No, and % immature Ret. Correlations improved for all 40 samples. The potential of %pRBC to predict high-risk patients was much higher (AUC=0.969, p < 0.0001) than that of %Fcell (AUC=0.873, p = 0.0019), with a sensitivity of 94.4% and a specificity of 88.9%, for %pRBC<65%.

The identification of samples with similar HbF% and distinct %pRBC (and vice-versa) highlights the unique ability of our qFC method to assess cHbF distribution in the RBC population of these patients. Evaluating consecutive samples from 7 patients, %pRBC<40% in 3 patient samples correctly anticipated an imminent clinical worsening, despite normal %Fcell and hematimetric indices.

Our results demonstrate an improved ability of the qFC method to stratify SCD patients, strongly suggesting it potential as a marker to predict patients at risk of vaso-occlusive events.