Posterior reversible encephalopathy syndrome (PRES) has been known as a neurological complication in patients after allogeneic hematopoietic stem cell transplantation (allo-HSCT). PRES is an acute or subacute syndrome with various neurological symptoms such as headache, encephalopathy, visual disturbance and seizures [1]. Neuroimaging may show subcortical white matter vasogenic edema typically in bilateral parieto-occipital lobes [1]. The etiology of PRES is not clear, which has been speculated as cerebrovascular dysregulation associated with exposure to toxic agents such as chemotherapeutic drugs and immunosuppressive agents [1]. Several studies have indicated that severe hypertension, Pesaro class III, and acute graft-versus-host disease (aGVHD) are potential risk factors for PRES in patients with transfusion-dependent thalassemia (TDT) after allo-HSCT [2, 3]. Severity of brain magnetic resonance imaging (MRI) findings and the length of in-hospital stay were independent risk factors related to adverse outcomes from PRES in patients with TDT after allo-HSCT [4].

We enrolled consecutive patients with TDT after allo-HSCT from October 2010 and May 2023 and assessed their clinical factors and brain MRI findings. Univariate and multivariate binary logistic regression were used to identify the risk factors for the occurrence and the prognosis of PRES. Receiver operating characteristic (ROC) curve analysis was performed to obtain the area under the curve (AUC) values for assessing the prediction efficacy. The study methods are detailed in the Supporting Information.

A total of 770 patients diagnosed with TDT underwent allo-HSCT during the study interval (Table S1). All patients underwent one allo-HSCT, with the mean age at 7.36 years. Among them, 64 patients had PRES with an incidence of 8.3%. The median time of onset for PRES from the start of allo-HSCT was 77 days (ranging from 12 to 441 days). Ten patients who developed PRES had poor prognosis, with a score of 3–6 on the modified Rankin Scale (mRS) clinical outcome scale (Supporting Information: Methods). Three of the 64 patients with PRES died, resulting in a mortality rate of 4.69%. Among the cohort of 64 cases with PRES, 45 cases showed mild brain MRI findings, 16 cases showed moderate MRI findings, and 3 cases exhibited severe MRI abnormalities. We found 6 patients showing hyperintensity on diffusion-weighted images (DWI b values = 1000s/mm²) and hypointensity in apparent diffusion coefficient (ADC). A summary of MRI findings of PRES is presented in Figure 1. The median duration for complete resolution of lesions during follow-up was 14 days (range 6–99 days). The median interval for lesion improvement was 21 days (range 3–66 days), with the earliest documented improvement observed at 3 days post-diagnosis.

Regarding the development of PRES, univariate analysis showed that age, serum ferritin, hypertension, pneumonia after allo-HSCT, cytomegalovirus (CMV) infection, electrolyte disturbance, aGVHD, chronic GVHD (cGVHD), and tacrolimus (TAC) were significantly associated with the occurrence of PRES (Table S2). Multivariate logistic regression analysis showed that age (OR, 4.75; 95% confidence interval [95% CI]: 2.23–10.15; p < 0.001), serum ferritin (OR, 5.30; 95% CI: 2.23–12.62; p < 0.001), hypertension (OR, 6.50; 95% CI: 4.37–9.66; p < 0.001), CMV infection (OR, 4.45; 95% CI: 1.90–10.44; p = 0.001), and aGVHD (OR, 2.68; 95% CI: 1.56–4.59; p < 0.001) were associated with the occurrence of PRES (Table S3). Three predictors for PRES including hypertension (stage 1 and 2), age (> 10 years) and serum ferritin (> 5000 ng/mL) were extracted from the five variables from multivariate analysis through ROC curve analysis (Figure 1). The area under the curve (AUC) values from the ROC curve analysis for hypertension, age and serum ferritin were 0.87 (95% CI: 0.81–0.93), 0.71 (95% CI: 0.65–0.77), and 0.70 (95% CI: 0.63–0.77) respectively. The AUC values for aGVHD and CMV infection were 0.69 (95% CI: 0.61–0.77) and 0.68 (95% CI: 0.61–0.76), respectively. The cut-off points for the risk factors identified were as follows: hypertension (1.5, sensitivity: 0.695, specificity: 0.960), age (2.5, sensitivity: 0.452, specificity: 0.826), serum ferritin (0.5, sensitivity: 0.678, specificity: 0.727), acute GVHD (aGVHD) (0.5, sensitivity: 0.492, specificity: 0.856), and CMV infection (0.5, sensitivity: 0.542, specificity: 0.840).

Regarding prognosis of PRES, univariate analysis showed that age, serum ferritin, MRI severity, CMV infection, pneumonia after allo-HSCT, and TAC were significantly associated with PRES prognosis (Table S4). Multivariate logistic regression analysis showed that MRI severity (OR, 6.749; 95% CI: 1.41–32.41; p = 0.017) was the independent risk factor for the prognosis of PRES (Table S5).

This study identified risk factors such as hypertension, age (> 10 years), and serum ferritin being associated with the occurrence of PRES. Hypertension has been emphasized as common adverse events of calcineurin inhibitors and has been associated with PRES. Similarly, a study has shown that severe hypertension is the independent risk factor of PRES post-HSCT in TDT [2]. The mechanism for hypertension being associated with PRES is not clear. Evidence suggests that dysregulation of cerebral blood flow may play a role in the development of PRES [1]. Rapidly developing hypertension in PRES may surpass cerebral blood flow autoregulation, which may cause hyperperfusion and breakdown of blood–brain barrier, allowing the interstitial extravasation of plasma and macromolecules [1]. Notably, the cerebral blood flow autoregulation threshold is lower in children than in adults, resulting in a lower mean blood pressure for PRES symptoms in pediatric patients [1]. To mitigate neurotoxicity during the transplantation process, it is advisable to maintain blood pressure close to baseline levels or lower when using calcineurin inhibitors. We identified that older age (> 10 years) was an independent risk factor for development of PRES post allo-HSCT in pediatric patients with TDT. A previous study suggested that patients over 10 years of age suffering from pediatric hematologic diseases and PRES were likely to have a poor outcome [5]. In this study, serum ferritin level was identified as a risk factor for development of PRES in patients with TDT after allo-HSCT, which was a novel finding that had not been reported in literature. Serum ferritin is a well-known inflammatory marker. In patients with TDT, excess iron catalyzes the formation of free radicals, exacerbating oxidative stress. This oxidative stress is a significant factor in the pathophysiology of TDT, contributing to cellular damage and inflammation. The inflammatory response prompts endothelial cells to release inflammatory cytokines, which leads to increased endothelial permeability and subsequently results in vascular edema [6]. Therefore, monitoring serum ferritin levels (< 5000 ng/mL) may play an important role in preventing PRES.

Our study also identified two risk factors such as MRI severity and hypointensity on ADC maps on brain MRI being predictors for poor prognosis in patients with TDT who had PRES after allo-HSCT, which was consistent with literature. Deng et al. similarly reported that MRI severity was an independent risk factor for poor prognosis from PRES [4]. Hypointensity on ADC map has been reported to be related to bad prognosis in patients with PRES [1]. High DWI signal intensity and pseudonormalized ADC values have been shown to be associated with cerebral infarction and may represent the earliest sign of non-reversibility as severe vasogenic edema progressing to cytotoxic edema [1].

The optimal timing for follow-up neuroimaging to monitor the recovery of PRES remains uncertain. In our cohort, the earliest time for the resolution of brain lesions upon follow-up was 3 days, with a median time of 14 days for complete resolution. The median time to observe lesion improvement was 21 days. Based on our findings and previous studies, we would suggest that a follow-up MRI should be acquired within 5–7 days after the initial scan, which may assist clinical decision making regarding therapeutic options.

In summary, we identified three risk factors including hypertension, older age (> 10 years) and high serum ferritin level for development of PRES and two risk factors such as high MRI severity and hypointensity on ADC maps for poor prognosis of PRES in the largest cohort to date of patients with TDT and PRES after allo-HSCT. Early diagnosis and prompt management of PRES could improve the outcome of the pediatric patients with TDT.

后部可逆性脑病综合征 （PRES） 被认为是同种异体造血干细胞移植 （allo-HSCT） 后患者的神经系统并发症。PRES 是一种急性或亚急性综合征，伴有各种神经系统症状，如头痛、脑病、视力障碍和癫痫发作 [1]。神经影像学检查可能显示皮质下白质血管源性水肿，通常发生在双侧顶枕叶[1]。PRES 的病因尚不清楚，据推测是与化疗药物和免疫抑制剂等毒性药物相关的脑血管失调 [1]。几项研究表明，严重高血压、Pesaro III 级和急性移植物抗宿主病 （aGVHD） 是同种异体 HSCT 后输血依赖性地中海贫血 （TDT） 患者发生 PRES 的潜在危险因素 [2， 3]。脑部磁共振成像 （MRI） 结果的严重程度和住院时间是与同种异体 HSCT 后 TDT 患者 PRES 不良结局相关的独立危险因素 [4]。

我们连续招募了 2010 年 10 月至 2023 年 5 月的 allo-HSCT 后 TDT 患者，并评估了他们的临床因素和脑部 MRI 结果。采用单因素和多因素二元 logistic 回归确定 PRES 发生和预后的危险因素。进行受试者工作特征 （ROC） 曲线分析，获得曲线下面积 （AUC） 值，用于评估预测效果。支持信息中详细介绍了研究方法。

共有 770 名诊断为 TDT 的患者在研究间隔期间接受了 allo-HSCT （表 S1）。所有患者均接受了 1 次同种异体造血干细胞移植，平均年龄为 7.36 岁。其中，64 例患者患有 PRES，发生率为 8.3%。从同种异体 HSCT 开始，PRES 的中位发病时间为 77 天 （范围从 12 天到 441 天不等）。10 例发生 PRES 的患者预后较差，改良 Rankin 量表 （mRS） 临床结果量表评分为 3-6（支持信息：方法）。64 名 PRES 患者中有 3 名死亡，死亡率为 4.69%。在 64 例 PRES 病例队列中，45 例出现轻度脑部 MRI 结果，16 例出现中度 MRI 结果，3 例出现严重 MRI 异常。我们发现 6 例患者在弥散加权图像上表现出高信号 （DWI b 值 = 1000s/mm ² ） 和表观弥散系数 （ADC） 低信号。图 1 总结了 PRES 的 MRI 结果。随访期间病灶完全消退的中位持续时间为 14 天 （范围 6-99 天）。病变改善的中位间隔为 21 天 （范围 3-66 天），最早记录的改善是在诊断后 3 天观察到的。

关于 PRES 的发展，单变量分析显示，年龄、血清铁蛋白、高血压、异种血干细胞移植后肺炎、巨细胞病毒 （CMV） 感染、电解质紊乱、aGVHD、慢性 GVHD （cGVHD） 和他克莫司 （TAC） 与 PRES 的发生显著相关（表 S2）。多变量logistic回归分析显示，年龄（OR，4.75;95%置信区间[95% CI]：2.23-10.15;p < 0.001）、血清铁蛋白（OR，5.30;95% CI：2.23-12.62;p < 0.001）、高血压（OR，6.50;95% CI：4.37-9.66;p < 0.001）、CMV 感染（OR，4.45;95% CI：1.90-10.44;p = 0.001）和 aGVHD（OR，2.68;95% CI：1.56-4.59;p < 0.001）与 PRES 的发生相关（表 S3）。PRES 的三个预测因素包括高血压（1 期和 2 期）、年龄（> 10 岁）和血清通过 ROC 曲线分析从多变量分析的 5 个变量中提取铁蛋白 （> 5000 ng/mL）（图 1）。高血压、年龄和血清铁蛋白的 ROC 曲线分析的曲线下面积 （AUC） 值为 0.87 （95% CI： 0.81–0.93）、0.71 （95% CI： 0.65–0.77） 和 0.70 （95% CI： 0.63–0.77）。aGVHD 和 CMV 感染的 AUC 值分别为 0.69 （95% CI： 0.61-0.77） 和 0.68 （95% CI： 0.61-0.76）。确定的危险因素的临界点如下：高血压 （1.5，敏感性： 0.695，特异性： 0.960），年龄 （2.5，敏感性： 0.452，特异性： 0.826），血清铁蛋白 （0.5，敏感性： 0.678，特异性： 0.727），急性 GVHD （aGVHD） （0.5，敏感性： 0.492，特异性： 0.856） 和 CMV 感染 （0.5，敏感性： 0.542，特异性： 0.840）。

关于 PRES 的预后，单因素分析显示年龄、血清铁蛋白、MRI 严重程度、CMV 感染、同种异体 HSCT 后肺炎和 TAC 与 PRES 预后显著相关（表 S4）。多变量 logistic 回归分析显示，MRI 严重程度 （OR， 6.749;95% CI： 1.41–32.41;p = 0.017） 是 PRES 预后的独立危险因素（表 S5）。

本研究确定了高血压、年龄 （> 10 岁） 和血清铁蛋白等危险因素与 PRES 的发生相关。高血压被强调为钙调神经磷酸酶抑制剂的常见不良事件，并与 PRES 相关。同样，一项研究表明，严重高血压是 TDT 中 HSCT 后 PRES 的独立危险因素 [2]。高血压与 PRES 相关的机制尚不清楚。有证据表明，脑血流失调可能在 PRES 的发展中发挥作用 [1]。PRES 患者快速发展的高血压可能超过脑血流自动调节，这可能导致血脑屏障过度灌注和破坏，导致血浆和大分子间质外渗 [1]。值得注意的是，儿童的脑血流自动调节阈值低于成人，导致儿科患者出现 PRES 症状的平均血压较低 [1]。为了减轻移植过程中的神经毒性，建议在使用钙调磷酸酶抑制剂时将血压保持在接近基线水平或更低水平。我们确定高龄 （> 10 岁） 是 TDT 儿科患者同种异体 HSCT 后发生 PRES 的独立危险因素。先前的一项研究表明，10 岁以上患有儿科血液病和 PRES 的患者可能会有不良预后 [5]。在这项研究中，血清铁蛋白水平被确定为同种异体 HSCT 后 TDT 患者发生 PRES 的危险因素，这是一个未在文献中报道的新发现。血清铁蛋白是一种众所周知的炎症标志物。在 TDT 患者中，过量的铁会催化自由基的形成，加剧氧化应激。 这种氧化应激是 TDT 病理生理学中的一个重要因素，导致细胞损伤和炎症。炎症反应促使内皮细胞释放炎性细胞因子，导致内皮通透性增加，随后导致血管水肿 [6]。因此，监测血清铁蛋白水平 （< 5000 ng/mL） 可能在预防 PRES 中起重要作用。

我们的研究还确定了两个危险因素，例如 MRI 严重程度和脑部 MRI 上 ADC 图上的低信号是同种异体 HSCT 后 PRES 的 TDT 患者预后不良的预测因素，这与文献一致。邓等人同样报道，MRI 严重程度是 PRES 预后不良的独立危险因素 [4]。据报道，ADC 图谱上的低信号与 PRES 患者的不良预后有关 [1]。高 DWI 信号强度和假正常化 ADC 值已被证明与脑梗死相关，并且可能是不可逆性的最早迹象，即严重的血管源性水肿进展为细胞毒性水肿 [1]。

随访神经影像学监测 PRES 恢复的最佳时机仍不确定。在我们的队列中，随访时脑病变最早消退时间为 3 天，完全消退的中位时间为 14 天。观察病灶改善的中位时间为 21 天。根据我们的发现和以前的研究，我们建议应在初次扫描后 5-7 天内进行随访 MRI，这可能有助于临床做出有关治疗方案的决策。

总之，我们确定了 3 个危险因素，包括高血压、高龄 （> 10 岁）和高血清铁蛋白水平，以及 2 个危险因素，例如 MRI 严重程度高和 ADC 图上的低信号，导致 PRES 预后不良，这是迄今为止最大的 TDT 和 PRES 患者队列同种异体 HSCT 后。PRES 的早期诊断和及时管理可以改善 TDT 患儿患者的预后。