This is a retrospective multicenter national study (ClinicalTrials.gov identifier NCT03397017). Data from the clinical records of eligible patients were collected from 2015 to 2018 through a standard web-based application (www.SITE-italia.org) encrypted by the Central Server. The recording of personal, therapy, and complications data of patients with sickle cell anemia included in the National Transfusion Treatment Survey was undertaken by the responsible investigator or sub-investigators selected by each center, following registration on the site. The operator can subsequently access the patient's clinical data and perform updates to follow clinical evolution over time. The study did not involve any additional tests compared to routine patient management.

The study aimed to identify the lifelong therapeutic approaches, focusing on transfusion regimens, in clinical management of a large multi-ethnic Italian cohort of patients with SCD. The study was coordinated by SITE, in collaboration with SIMTI and AIEOP. A total of 34 Italian reference centers from 15 Italian regions were involved. Data were collected through the National Comprehensive Reference Centers for SCD. The study was approved by the Ethics Committee of Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico of Milan, Italy, the coordinating center of the study.

The inclusion and exclusion criteria were, respectively, the presence or absence of a biochemical and/or genetic diagnosis of SCD and the capability or not to give informed consent. Written informed consent was obtained from all included patients after the study aims and protocol had been explained. When necessary, a Cultural Linguistic Mediator could be engaged to assist in the consent process. Besides demographic data, every center collected data concerning the type and sequence of therapy, acute transfusion therapy (ATR), chronic transfusion therapy (CTR, i.e., regular transfusion regimen for more than 1 year), or HC, the main indications for ACR and CTR, and details on alloimmunization (number and systems of antibodies) over the previous 4 years of treatment.

Descriptive statistics of continuous variables were expressed using medians and interquartile range (IQR: 25th-75th percentile); dichotomous variables were summarized as counts and percentages. The Shapiro-Wilk test was used to test the normality of distribution. As no normal distribution was found, the differences between groups were analyzed with the Wilcoxon signed-rank test. Two- or multiple-proportion comparisons between variables were performed with the Z-test. All tests were two-sided, and p-values < 0.05 were considered as statistically significant. Data management and graphs were created with Excel (Microsoft, Seattle, WA, USA), and statistical analysis was performed using R version 3.5.1 for Windows (R Core Team, Vienna, Austria).

Our study shows that Italian SCD patients overall cluster into two distinct groups: children of African and adults of Caucasian descent (Figure 1A). We collected data from 1,579 patients, of whom 802 were male (51%) and 777 were female (49%) with a median age of 23 (IQR 10–41) years (Supplementary Figure S1). No significant difference in age was found between males and females (p = 0.18); the median age was 21.5 (IQR 10–40) years in males and 25.3 (IQR 10–42) years in females. Ethnicity was equally distributed between Caucasians (47.9%) and Africans (48.7%), with a minority of African American (2.8%) and Asian (0.4%) SCD patients. Ethnicity was not defined for 0.2% of patients (Table 1). In African patients, the most common genotype was homozygosity for HbS (SS; 71.3%). Compound heterozygosity with β-thalassemia mutations was 2.3% for S/β0-thalassemia, 3.9% for S/β+-thalassemia, and 20.8% for SC disease. Otherwise, the most prevalent genotype in Caucasians was S/β-thalassemia (74.5%; in particular, S/β⁰-thalassemia 34.8% and S/β⁺-thalassemia 39.7%). SS genotype was present in 21.4% and SC in 0.5% of SCD patients of Caucasian descent. The median age of African SCD patients was 12 (IQR 7–22) years compared with 38 (IQR 25–49) years in Caucasian patients (p < 0.001).

Full data on therapeutic management of SCD patients were collected for 1,364 out of the 1,579 patients (Table 2; Figure 1B; Supplementary Figure S1). The following therapeutic regimens were identified: (i) HC: chronic hydroxycarbamide therapy for more than 1 year; (ii) CTR: chronic transfusion regimen; (iii) ATR: acute transfusion approach; and (iv) no treatment (Figure 2A). When we considered therapeutic strategies in function of age, we found that blood transfusion approaches were used throughout patient's journey (Figure 2B). A total of 292 patients (21.4%) were on CTR, 196 (14.4%) on ATR, and 497 on HC (36.4%), of whom 100 needed transfusions for acute VOCs. Three hundred and seventy-nine (27.8%) patients never received long-term therapy. In both ATR and CTR clinical setting, donor-recipient exact matching for serological ABO, Rhesus and Kell antigen for RBC compatibility was carried out according to international and Italian guidelines, released in 2014 (9, 13, 24–27).

The main indications for CTR were recurrent severe VOCs (371 events; 37.1%), chronic anemia (306 events; 30.6%), primary stroke prevention (78 events, 7.8%) and secondary stroke prevention (55 events, 5.5%), acute chest syndrome (ACS; 107 events; 10.7%), splenic sequestration (31 events, 3.1%), HC-resistant pain (39 events; 3.9%) and leg ulcers (12 events; 1.2%) (Figure 2B). The median value of hemoglobin (Hb) before CTR was 9.2 g/dL (IQR: 8.5–10, min-max: 5.2–14.9 g/dL, data available for 85% of patients) and the median value of HbS was 50% of total Hb (IQR: 39–55, min-max: 8.5–89%, data available for 67% of patients).

The main indications for ATR were acute anemia (384 events; 35%), VOCs (352 events; 32.1%), ACS (170 events; 15.5%), surgery (82 events; 7.5%), pregnancy (64 events; 5.8%), splenic sequestration (26 events; 2.4%), stroke (9 events; 0.8%), multi-organ failure (6 events; 0.5%) and priapism (5 events; 0.5%) (Figure 2B). Before ATR, the median value of Hb level was 8.0 g/dL (IQR: 6.8–9.0 g/dL, min-max 2.7–15.2 g/dL, data available for 72% of patients), with 67% of HbS (IQR: 56–78%, min-max: 30–93, data available for 43% of patients). Patients who were not treated had a median Hb value of 10.7 g/dL (IQR: 9.3–12 g/dL). When we stratified SCD population according to age: children (<18 years) and adults (>18 years), the proportion of different indications for ATR was similar within the two groups except for pregnancy and priapism, which were present in adults (Supplementary Figure S2A). Considering CTR, we found CTR to be more frequently introduced in adults than in children with SCD (Supplementary Figure S2A). CTR protocols were more frequently applied in adults than in children to prevent ACS and to treat chronic pain resistant to HC therapy (Supplementary Figure S2A). We found a higher proportion of Caucasian adults (older than 20 years) under CTR (135/181, 75%) when compared to African adults (41/94, 44%). In children and adolescents (younger than 20 years), the proportion of Caucasians treated with CTR (46/181, 25%) was lower when compared with Africans (53/94, 56%). The ratio between patients in ATR and CTR decreases with during aging (p < 0.05, Figure 2A).

HC was generally started after 5 years of age in SCD patients of both Caucasian (6%) and African (94%) descent. Two hundred and fourteen of 246 Caucasian patients were adults (17% SS and 79% Sβ genotypes). As expected, SCD patients with SC genotype showed less intensive medical treatment than SCD patients with either SS or Sβ^0/+ genotypes (Figure 1B).

For 250 out of 1,364 patients (18.3%), it was possible to follow the temporal sequence of change in therapy (Table 2). Fourteen patients (1.0% of the 1,364) switched from HC to CTR, and 236 patients (17.3%) switched from CTR to HC. A subset of 60 patients (4.3% of the 1,364 with treatment data available) received a simultaneous blood transfusion and HC for a median period of 4.6 years (IQR 2.2–11.1 years). This was due to the severity of sickle cell-related clinical presentation, which was only partially controlled by HC therapy. The main indications for CTR in these 60 patients were, respectively, VOCs (number of events: 33), anemia (n = 25), ACS (n = 13). HC-resistant pain (n = 6), splenic sequestration (n = 4), primary (n = 2) or secondary (n = 3) stroke prevention. Combined therapy (CTR+HC) was started at a median age of 20.5 (IQR 10.7–30.8) years. In 14/250 patients, HC was started at a median age of 30.7 (IQR 28.1–41.4) years, and CTR needed to be added to HC after a mean period of 3.8 (IQR 2.4–5.3) years at a median age of 34.4 (IQR 30.3–48.3) years. The proportion of SCD patients without any treatment decreased with aging (Figure 2A). Two out of 1,364 patients with SCD displayed hyperhemolysis reaction after RBC transfusion (1 African SS child: 5 years of age; 1 African SS adult: 23 years of age).

Collectively, our data indicate that acute and chronic transfusion approaches were chosen to limit disease progression in patients of both Caucasian and African descent. RBCs from Caucasian donors were used. The donor-recipient exact matching for ABO, Rhesus and Kell antigen for RBC compatibility strategy limits alloimmunization, which might further complicate clinical management of patients with SCD (9, 18, 19, 28).

In our study, alloimmunization was documented in 135 (8.5%) out of 1,579 patients with SCD. Sixty-one of these 135 patients (45.2%) were already alloimmunized at their arrival at the comprehensive SCD center. Four of them (6.6%) developed alloimmunization against new RBC antigen (Supplementary Figure S3A) 36 of these 135 patients (27%) developed alloantibodies after they arrived at the Italian referral center (Supplementary Figure S3A). Data on alloantibody status in relation to the timing of arrival at the SCD center were not available for the other 38 alloimmunized SCD patients.

Among the alloimmunized SCD patients, 33 (24.4%) were younger than 18 years, 53 (39.3%) were aged 18–40 years, and 49 (36.3%) were older than 40 years (Supplementary Figure S3B). Fifty-one alloimmunized patients were African [median age 20 years (IQR 11–28 years)], corresponding to 6.6% of the African group; 78 patients were Caucasian [median age 41 years (IQR 31–51 years)], representing 10.3% of the Caucasian group; five patients were African American [median age 35 years (IQR 30–37 years)] and for one patient (age 29 years) ethnicity was not reported. Thus, alloimmunization was more frequent in adults than in children with SCD (p < 0.001). The frequency of alloimmunization was similar in female and male patients with SCD.

RBC-specific antibodies were detected in 124 patients, of whom 70 (41 males, 29 females) had a single antibody and 43 (16 males and 27 females) had multiple antibodies (59.5%, 28.6, 9.5, and 2.4%, respectively, with two, three, four, and six antibodies). Specific antibodies were not identified in 11 patients (3 males and 8 females) (Supplementary Figure S3C). A higher proportion of Caucasian patients had multiple antibodies compared with patients of other ethnicities. In the Caucasian group, a single antibody was prevalently found in males (24/40, 60%), while there were no significant differences between genders in the African group. When genotypes were considered, we observed alloimmunization to be more frequently associated with SS genotype in both African and African American patients, whereas SS and Sβ^0/+ genotypes were equally associated with alloimmunization in SCD patients of Caucasian descent (Supplementary Figure S3B).

As shown in Figure 3, alloimmunization predominantly involves RhCDE and Kell systems, with Rhesus (45%) the most represented antibody system, followed by Kell (15%), MNS (12%), Duffy (6.2%), Lewis (4.8%), Kidd (4.3%) and Lutheran (2.7%). In addition, antibodies against minor antigens (0.8%) and antibodies not better identified (8.8%) were also detected. Of interest, anti-e antibodies were found in two e-negative African SCD patients. When we considered alloimmunization profile and patients' age, a higher rate (26 vs. 6%) of alloantibodies with non-identified specificity in children's group was observed (Figure 3, right pie-charts).

Eleven patients (8.1%) with alloimmunization developed a delayed hemolytic transfusion reaction (DHTR) (8 females and 3 males) with a median age of 39 years (IQR: 18.5–49). Six of these patients had SS genotype (4 Africans, 2 Caucasians), and five were Caucasian with Sβ genotype. In 4 out of the 11 patients, a single antibody (D, C, Cw, Jsb) was detected, and in another 4 of the 11 (all females, 3 Caucasians, and 1 African) multiple antibodies (2–4 different specificities) were detected. In 2 patients, antibodies were not clearly identified. Taken together, these data indicate that low alloimmunization events recorded in both African and Caucasian patients might most likely be related to the donor-recipient exact matching for ABO, Rhesus and Kell antigen for RBC compatibility.