The implementation of newborn screening associated with comprehensive care has led to the transformation of sickle cell disease (SCD) from a fatal illness in children to a chronic disease with multiple organ dysfunctions and premature death in adults.^{1, 2} Organ injury begins during childhood, as a result of chronic hemolysis-related endothelial dysfunction and hypoxia triggered by repeated vaso-occlusive crisis (VOC). In our recently published prospective neonatal cohort study, the first signs of extracerebral major organ impairment were detected later than acute SCD-related events. However, our attempts to follow the progression of these conditions from adolescence to early adulthood were hampered by a median duration of follow-up (FU) of only 13 years.³

Our main objective is now to describe cumulative changes in risk for extracerebral chronic organ damage between the ages of 15 and 25 years. We, therefore, selected all patients born before October 2003 with a theoretical age of at least 18 years, regardless of their vital status, from our regional longitudinal newborn cohort in Creteil Hospital, and retrospectively collected clinical data for these patients after their transfer to adult SCD referral centers in France. Kaplan–Meier (KM) survival and event-free survival estimates and 95% confidence intervals (95%CI) were calculated for all SCD-related complications, and first use of disease-modifying therapies (DMT) [hydroxyurea (HU), transfusion program (TP), or hematopoietic stem cell transplantation (HSCT)]. Our secondary objective was to identify early predictive factors for chronic organ injury, by analyzing baseline blood parameters recorded during early childhood, before the introduction of DMT. We limited this analysis to adult patients with a last visit recorded after the age of 15 years, and we compared the groups of patients with and without each organ complication. We analyzed the HbSS/HbSβ⁰ subgroup, the HbSC subgroup being too small for analysis.

Our newborn cohort includes 150 subjects, 107 with HbSS and six with HbSβ⁰ (these two groups were pooled into a single group on the basis of their known clinical similarity), 10 with HbSβ⁺ and 27 with HbSC sickle genotypes, with a mean (±SD) duration of FU of 20.7 ± 7 years. Demographic and baseline biological characteristics are summarized, together with FU data and the prevalence of chronic organ complications, in Table S1A.

Two deaths occurred after the age of 15 years, both in the HbSS genotype group. One boy died at the age of 19 years from chronic graft-versus-host disease after haploidentical HSCT, and the other died at the age of 28 years, during an episode of acute chest syndrome (ACS). Both had chronic organ damage (cholangiopathy for the younger patient and nephropathy, dilated cardiomyopathy, and pulmonary hypertension for the older patient). Six children died before the age of 15 years. None of these children had any signs of chronic organ complications at death. Consistent with other reports, the probability of survival at 25 years was 94.9% [95% CI: 90.5%–98%] for the total study population.⁴

The initiation of DMT over time is shown in Figure S1A. The overall probability of first HU use increased from 49.1% [95%CI: 39.4%–58.8%] at 15 years to 68.3% [95%CI:57.3%–78.3%] at 25 years for the severe genotype group. HU treatment was initiated after the age of 15 years in 21 HbSS/Sβ⁰ patients, in accordance with French national recommendations: for recurrent acute vaso-occlusive complications (sixteen), severe chronic anemia (two), osteonecrosis (one) and as a substitute for TP in two children with stroke.

A first TP was initiated after the age of 15 years in only eight adolescents, all from the HbSS/Sβ⁰ group: due to persistent VOC/ACS in three, the prevention of silent stroke recurrence in two, an increase in tricuspid regurgitation velocity (TRV) ≥2.7 m/s in two and in preparation for HSCT in one.

In the HbSC genotype subgroup, phlebotomy was the most common therapeutic intervention, initiated for 10 patients: to reduce dizziness/headache symptoms in four, to prevent VOC recurrence in four, and for proliferative retinopathy in two. Two patients with Sβ⁺ genotype started phlebotomy because of dizziness/headache symptoms. In four patients with the HbSS/Sβ⁰ genotype, phlebotomy was performed in association with HU treatment, to reduce viscosity or iron overload.

The cumulative risk of major extracerebral organ complications—high TRV, albuminuria, retinopathy restricted to the most severe injury (proliferative retinopathy and use of laser photocoagulation), and symptomatic avascular osteonecrosis—is shown in Figure 1. The long-term risks of cholelithiasis, cholecystectomy, and splenectomy are shown in Figure S1B.

We considered TRV to be high for values ≥2.7 m/s, as this threshold has recently been shown to be appropriate for use in adolescents and young adults because significantly associated with death. The overall probability of a high TRV increased from 2.5% [95% CI: 0.5%–6%] at 15 years to 13% [95% CI:6.6%–21.1%] at 25 years, for all genotype groups. For the identification of risk factors for a high TRV, we compared the group of nine adolescents with a high TRV (≥2.7 m/s), with the remaining 98 adolescents with a normal TRV. In univariate analysis, high TRV was significantly associated with higher platelet counts and lower Hb and HbF levels at baseline. Alpha-thalassemia (2 or 3/4 genes) was highly protective (Table S1B). Low baseline HbF levels remained independently associated with high TRV in multivariate analysis.

We evaluated sickle cell nephropathy by assessing albuminuria (defined as an albumin-to-creatinine ratio ≥ 30 mg/g) as an indicator of early glomerular damage contributing to the development and progression of chronic kidney disease and associated with death in adults.¹ The risk of microalbuminuria was lower in this cohort than in previous studies,⁵ but increased from 6% [95% CI: 2.6%–10.7%] at 15 years of age to 20.8% [95% CI: 12.8%–30.1%] at 25 years of age, with a significant difference between the severe (25.3% [95% CI: 15.4%–36.9%]) and HbSC (4.3% [95% CI: 0.16%–2%]) genotypes. Five children (all HbSS/Sβ⁰) developed macroalbuminuria, at the ages of 6 (in association with renal artery stenosis), 21, 23, 26, and 28 years. For the identification of risk factors for albuminuria, we compared the group of 23 adolescents with albuminuria to the remaining 77 adolescents without albuminuria. In univariate analysis, high bilirubin levels and mean corpuscular volumes were associated with a higher risk of albuminuria, confirming the contribution of excess hemolysis to albuminuria.^{2, 6, 7} Again, higher HbF levels at baseline were protective (Table S1B). Albuminuria was significantly correlated with a higher risk of high TRV (whether defined as ≥2.5 or ≥2.7 m/s).

We assessed ophthalmic complications by focusing on proliferative retinopathy, and severe retinopathy requiring laser photocoagulation. Our findings confirm that the frequency of severe retinopathy increases from adolescence (from 1.5% to 17% between the ages of 15 and 25 years) and is much higher in patients with the HbSC genotype than in those with severe genotypes (probability of 36.1% [95% CI: 15.3%–60.1.3%] by 25 years of age versus 7.1% [95% CI: 2.2%–14.6%] for severe genotypes). Both non-proliferative and proliferative retinal manifestations are caused by erythrostasis secondary to sickling. The basis of the association between HbSC disease and retinopathy remains unclear, but its existence confirms the need for earlier screening in children with HbSC disease, and regular screening in all adolescents with SCD.

Eighteen children were diagnosed with symptomatic avascular osteonecrosis, affecting the hips in all but two cases. Seventeen of these children were in the severe genotype group. The probability of symptomatic osteonecrosis was significantly higher in the severe SCD genotype group, increasing from 6% [95% CI: 2.2–11.5] at 15 years of age to 19% [95% CI: 10.2%–29.6%] by the age of 25 years. Avascular osteonecrosis was associated with age, but not with any of the baseline biological parameters.

Albuminuria, high TRV, and proliferative retinopathy were not detected after HSCT, in any of the 28 patients who underwent transplantation. Considering the severe genotype group only and at the time the organ complication was detected: among the 23 patients who developed albuminuria, 15 were not receiving any DMT (7 had discontinued DMT from 3 months to 20 years (mean 5 years ±7 SD) before albuminuria diagnosis) but 8 were taking HU. Among the 10 patients who developed high TRV, 3 were not receiving any DMT, but HU therapy was ongoing for 6 patients and TP for one patient. Finally, among the 6 patients who were diagnosed with proliferative retinopathy, 5 were taking HU.

The probability of cholecystectomy increased steadily from early childhood to early adulthood, with no plateau, reaching a value close to the cumulative risk of cholelithiasis of 54.9% [95% CI: 44.9%–64.6%] at 25 years. Splenectomy was performed only in patients with the severe genotype and early in childhood (median age of 5.3 years [2.8–13.8 years]). None of the patients underwent splenectomy after the age of 15 years.

In summary, we report a high risk of progression for chronic extracerebral major organ complications of SCD throughout adolescence and early adulthood, a period during which there is a risk of discontinuity of care and HU compliance difficulties. A focused “transition program” to ensure the safe transfer of teenagers from pediatric to adult services should, therefore, be a key priority, to facilitate screening for chronic organ dysfunction.

Our study highlights the considerable and persistent morbidity of SCD due to chronic organ damage when hydroxyurea treatment is initiated after the recurrence of vaso-occlusive complications. Our findings also confirm the predictive value for albuminuria and high TRV of biological markers recorded during early childhood, including markers of high hemolysis and low levels of HbF. Hydroxyurea is the primary and most well-established pharmacologic therapy with proven benefits to ameliorate the clinical course of SCD and reduce chronic hemolysis, primarily due to its ability to increase the expression of HbF, which prevents sickling of red blood cells, decreasing the likelihood of infarction and hemolysis-related organ damage. Our data suggest that the implementation of early preventive HU treatment, with escalation to the maximal tolerated dose, may help to prevent chronic organ damage in patients with HbSS/Sβ⁰ genotypes. Moreover, the detection of organ damage should lead to early genoidentical HSCT.

新生儿筛查与综合护理的实施已导致镰状细胞病（SCD）从儿童致命疾病转变为导致成人多器官功能障碍和过早死亡的慢性疾病。^{1, 2}器官损伤始于儿童时期，这是由于慢性溶血相关的内皮功能障碍和反复血管闭塞危机 (VOC) 引发的缺氧所致。在我们最近发表的前瞻性新生儿队列研究中，脑外主要器官损伤的最初迹象的检测晚于急性 SCD 相关事件。然而，我们追踪这些疾病从青春期到成年早期的进展的尝试因中位随访时间 (FU) 仅 13 年而受到阻碍。³

我们现在的主要目标是描述 15 岁至 25 岁之间脑外慢性器官损伤风险的累积变化。因此，我们从克雷泰伊医院的区域纵向新生儿队列中选择了所有 2003 年 10 月之前出生、理论年龄至少为 18 岁的患者，无论其生命状况如何，并回顾性收集了这些患者转为成人后的临床数据。法国的 SCD 转诊中心。针对所有 SCD 相关并发症以及首次使用疾病缓解疗法 (DMT) [羟基脲 (HU)、输血，计算 Kaplan-Meier (KM) 生存期和无事件生存期估计值以及 95% 置信区间 (95%CI)计划（TP）或造血干细胞移植（HSCT）]。我们的次要目标是通过分析 DMT 引入之前儿童早期记录的基线血液参数来确定慢性器官损伤的早期预测因素。我们将这项分析仅限于最后一次就诊记录在 15 岁之后的成年患者，并比较了有和没有每种器官并发症的患者组。我们分析了 HbSS/HbSβ ⁰亚组，HbSC 亚组太小，无法进行分析。

我们的新生儿队列包括 150 名受试者，其中 107 名患有 HbSS，6 名患有 HbSβ ⁰（根据已知的临床相似性，将这两组合并为一组），10 名患有 HbSβ ⁺，27 名患有 HbSC 镰状基因型，平均 ( ±SD) FU 持续时间为 20.7 ± 7 年。表 S1A 总结了人口统计学和基线生物学特征，以及 FU 数据和慢性器官并发症的患病率。

两例死亡发生在 15 岁之后，均属于 HbSS 基因型组。一名男孩在半相合 HSCT 后于 19 岁时死于慢性移植物抗宿主病，另一名男孩于 28 岁时死于急性胸部综合征 (ACS)。两人都患有慢性器官损伤（年轻患者患有胆管病，老年患者患有肾病、扩张型心肌病和肺动脉高压）。六名儿童在 15 岁之前死亡。这些儿童在死亡时都没有任何慢性器官并发症的迹象。与其他报告一致，整个研究人群 25 年生存概率为 94.9% [95% CI：90.5%–98%]。⁴

DMT 随着时间的推移而开始如图 S1A 所示。对于严重基因型组，首次使用 HU 的总体概率从 15 年时的 49.1% [95%CI：39.4%–58.8%] 增加到 25 年时的 68.3% [95%CI：57.3%–78.3%]。根据法国国家建议， 21 名 HbSS/Sβ ⁰患者在 15 岁后开始进行 HU 治疗：针对复发性急性血管闭塞并发症（16 例）、严重慢性贫血（2 例）、骨坏死（1 例）和作为在两名中风儿童中替代 TP。

仅 8 名青少年在 15 岁后开始首次 TP，全部来自 HbSS/Sβ ⁰组：由于 3 名青少年持续存在 VOC/ACS，2 名患者预防无症状中风复发，三尖瓣反流速度增加（ TRV) ≥2.7 m/s（其中 2 例），并为 HSCT（1 例）做准备。

在 HbSC 基因型亚组中，放血是最常见的治疗干预措施，对 10 名患者进行了治疗：减少 4 名患者的头晕/头痛症状，预防 4 名患者的 VOC 复发，以及 2 名患者的增殖性视网膜病变。两名 Sβ ⁺基因型患者因头晕/头痛症状开始放血。在 4 名 HbSS/Sβ ⁰基因型患者中，进行了放血联合 HU 治疗，以降低粘度或铁超负荷。

主要脑外器官并发症（高 TRV、蛋白尿、仅限于最严重损伤的视网膜病变（增殖性视网膜病变和激光光凝术）以及症状性缺血性骨坏死）的累积风险如图 1 所示。胆囊切除术和脾切除术如图 S1B 所示。

我们认为 TRV 值≥2.7 m/s 时较高，因为该阈值最近已被证明适合在青少年和年轻人中使用，因为与死亡显着相关。对于所有基因型组，高 TRV 的总体概率从 15 年时的 2.5% [95% CI：0.5%–6%] 增加到 25 年时的 13% [95% CI：6.6%–21.1%]。为了识别高 TRV 的危险因素，我们将 9 名具有高 TRV (≥2.7 m/s) 的青少年组与其余 98 名具有正常 TRV 的青少年进行了比较。在单变量分析中，高 TRV 与基线时较高的血小板计数和较低的 Hb 和 HbF 水平显着相关。 α-地中海贫血（2 或 3/4 基因）具有高度保护性（表 S1B）。在多变量分析中，低基线 HbF 水平仍然与高 TRV 独立相关。

我们通过评估白蛋白尿（定义为白蛋白与肌酐比率≥30 mg/g）作为早期肾小球损伤的指标来评估镰状细胞肾病，该损伤有助于慢性肾病的发生和进展，并与成人死亡相关。¹该队列中微量白蛋白尿的风险低于之前的研究⁵ ，但从 15 岁时的 6% [95% CI：2.6%–10.7%] 增加到 20.8% [95% CI：12.8%–30.1%] ] 25 岁时，重症（25.3% [95% CI：15.4%–36.9%]）和 HbSC（4.3% [95% CI：0.16%–2%]）基因型之间存在显着差异。 5 名儿童（均为 HbSS/Sβ ⁰）出现大量白蛋白尿，年龄分别为 6 岁（与肾动脉狭窄相关）、21 岁​​、23 岁、26 岁和 28 岁。为了确定白蛋白尿的危险因素，我们将 23 名患有白蛋白尿的青少年组与其余 77 名没有白蛋白尿的青少年进行了比较。在单变量分析中，高胆红素水平和平均红细胞体积与较高的蛋白尿风险相关，证实了过度溶血对蛋白尿的影响。^{2, 6, 7}同样，基线时较高的 HbF 水平具有保护作用（表 S1B）。蛋白尿与高TRV（无论定义为≥2.5还是≥2.7 m/s）的较高风险显着相关。

我们通过关注增殖性视网膜病变和需要激光光凝的严重视网膜病变来评估眼科并发症。我们的研究结果证实，严重视网膜病变的发生率从青春期开始增加（15 至 25 岁之间从 1.5% 增加到 17%），并且 HbSC 基因型患者的严重视网膜病变发生率远高于严重基因型患者（概率为 36.1%）。 25 岁时，95% CI：15.3%–60.1.3%]，而严重基因型则为 7.1% [95% CI：2.2%–14.6%]。非增殖性和增殖性视网膜表现都是由镰状细胞继发的红斑瘀滞引起的。 HbSC 疾病与视网膜病变之间关联的基础尚不清楚，但其存在证实需要对患有 HbSC 疾病的儿童进行早期筛查，并对所有患有 SCD 的青少年进行定期筛查。

18 名儿童被诊断患有有症状的股骨头坏死，除了 2 例外，所有儿童均影响髋部。其中十七名儿童属于严重基因型组。严重 SCD 基因型组出现症状性骨坏死的概率显着较高，从 15 岁时的 6% [95% CI: 2.2–11.5] 增加到年龄增长的 19% [95% CI: 10.2%–29.6%] 25 年。股骨头缺血性坏死与年龄相关，但与任何基线生物学参数无关。

在 28 名接受移植的患者中，HSCT 后均未检测到蛋白尿、高 TRV 和增殖性视网膜病变。仅考虑严重基因型组和检测到器官并发症时：在出现蛋白尿的 23 名患者中，15 名患者未接受任何 DMT（7 名患者在治疗前 3 个月至 20 年（平均 5 年±7 SD）已停止 DMT）。蛋白尿诊断），但 8 人服用 HU。在出现高 TRV 的 10 名患者中，3 名未接受任何 DMT，但 6 名患者正在进行 HU 治疗，1 名患者正在进行 TP。最后，在 6 名被诊断患有增殖性视网膜病变的患者中，有 5 名正在服用 HU。

从儿童早期到成年早期，胆囊切除术的概率稳步增加，没有出现平台期，在 25 岁时达到接近胆石症累积风险 54.9% [95% CI: 44.9%–64.6%] 的值。仅对具有严重基因型和儿童早期（中位年龄 5.3 岁 [2.8-13.8 岁]）的患者进行脾切除术。没有患者在 15 岁后接受脾切除术。

总之，我们报告整个青春期和成年早期 SCD 慢性脑外主要器官并发症进展的风险很高，在此期间存在护理中断和 HU 依从性困难的风险。因此，确保青少年从儿科服务安全转移到成人服务的重点“过渡计划”应该是一个关键优先事项，以促进慢性器官功能障碍的筛查。

我们的研究强调，当血管闭塞并发症复发后开始羟基脲治疗时，由于慢性器官损伤，SCD 的发病率相当高且持续存在。我们的研究结果还证实了儿童早期记录的白蛋白尿和高 TRV 生物标志物（包括高溶血标志物和低 HbF 水平的标志物）的预测价值。羟基脲是主要且最成熟的药物疗法，已被证明有利于改善 SCD 的临床病程并减少慢性溶血，这主要是因为它能够增加 HbF 的表达，从而防止红细胞镰状化，降低发生镰状细胞病的可能性。梗塞和溶血相关的器官损伤。我们的数据表明，实施早期预防性 HU 治疗，并逐步升级至最大耐受剂量，可能有助于预防 HbSS/Sβ ⁰基因型患者的慢性器官损伤。此外，器官损伤的检测应导致早期基因相同的 HSCT。