To the Editor:

In 2019, voxelotor was approved by the United States Food and Drug Administration for the treatment of sickle cell disease (SCD) in patients ≥16 years. The HOPE trial showed that treatment with voxelotor reduced hemolysis and improved anemia and rates of adverse and serious adverse events were no different between high dose, low dose, and placebo groups.¹ However, clinical trials do not provide all the information required for physicians to know how well the drug works in clinical practice. Postapproval real-world studies can identify barriers to use as well as risks and benefits not recognized in the initial clinical studies. We sought to examine voxelotor use in a real-world cohort of patients followed at a comprehensive urban sickle cell center. We hypothesized that economic and social barriers could prevent some patients from being able to obtain voxelotor, medication adherence may not be optimal, and that adverse events or side effects might result in fewer clinical and hematological benefits than those seen in the HOPE study.

Data were obtained retrospectively from electronic medical records (EMRs) at our center. Patients with SCD who had been prescribed voxelotor from November 25, 2019 to December 31, 2020 were included in this analysis. Use of hydroxyurea and erythropoiesis-stimulating agents (ESA) were defined by the presence of active prescription in the EMR. Demographics, clinical outcomes, and SCD-related morbidities were abstracted from the EMR.

Voxelotor therapy group included patients who verbally confirmed obtaining and starting the medication (VOX+), all others were defined as Vox−. Laboratory test results were obtained prior to ordering voxelotor, at steady state, defined as a visit with no acute complaints and at least 14 days from last episode of hospital or emergency room (ED) utilization. For VOX+ patients a second set of labs was obtained at least 14 days after starting voxelotor. Presence of voxelotor in patients' plasma was determined by changes in capillary zone electrophoresis based on methodology described in prior studies.^2-4 Episodes which resulted in a visit to the ED or admission to our hospital and number of packed red blood cell transfusions were noted for the 3 months prior to, and the 3 months post therapy initiation. Data on side effects, dose modifications, reasons for terminal discontinuation were collected.

Demographics, medications used, SCD-related morbidity, baseline ED visits, admissions, transfusion data, and baseline labs were compared between VOX+ and VOX− groups using a Mann Whitney U test for continuous variables and a Fisher's exact test for categorical variables. Baseline labs, ED visits, admissions and transfusions were compared to labs, ED visits, admissions, and transfusions post voxelotor using the Mann Whitney U test. Correlation of side effects with drug discontinuation and dose modification was examined using Fisher's exact test. Analyses were carried out using Stata software (version 15).

Between November 2019 and December 2020, 65 adults with HbSS/HbSβ0 were prescribed voxelotor; 57% female; mean age 40.7 ± 12.2 years old, 58% on hydroxyurea, 14% had received at least one red blood cell transfusion in the previous 3 months and 19% were on an ESA. Of the 65, 43 (66%) patients started the medication during the time of this study, at a mean of 48.4 days (SD 53.2 days) after it was prescribed (VOX+), while 22 (34%) never started (VOX−).

The most common reason for not starting (VOX−) was the inability to complete the application process for the drug, as it was dispensed from few specialty pharmacies with the assistance of Global Blood Therapeutics (GBT) care coordination staff (9/22, 41%). Other causes included lack of insurance approval (4/22, 18%), high copay (1/22, 5%), lost to follow-up (4/22, 18%), patient refusing therapy (1/22, 5%), decision to pursue an alternative therapy (2/22, 9%), and patient death (1/22, 5%).

Of the 43 VOX+ patients, 39 were started on a dose of 1500 mg, 3 were started at 1000 mg, and 1 was started at 500 mg. VOX− pts were more likely to have a known alpha thalassemia mutation, while VOX+ pts were more likely to have a history of lower extremity ulcers and higher baseline aspartate aminotransferase. All other characteristics were similar between VOX+ and Vox− patients (Table S1).

The median increase in hemoglobin was 1.5 g/dL and 22/33 (66%) showed an increase of 1 g/dL or more (Table 1, Figure S1). Laboratory data demonstrated a decrease in %HbF (p = .0018) but no change in total HbF (p = .35). This suggests the change in percent F is likely dilutional due to hemoglobin S surviving longer in circulation due to the reduction in hemolysis. A statistically significant increase in ALT (p = .0021) was observed, which had not been reported in the Hope trial. While none of the patients in this cohort required a dose modification for this, three recently seen patients developed Grade 2 (2) and Grade 3 (1) ALT abnormalities within 2 weeks after starting drug, all improved after either dose reduction or discontinuation. We therefore recommend monitoring liver function tests, and to expect a slight reduction in percent hemoglobin F after starting voxelotor. Voxelotor was detected in the plasma of 78% of patient who reported taking it. No change was seen in frequency of transfusions (p = .4), ED visits (p = .79) or admissions (p = .32).

At the time of this review, patients had been on voxelotor an average of 48 weeks. Of those 29/43 (67%) noted at least one side effect and 14/43(33%) noted multiple side effects. Of those who experienced side effects: 8/29 (28%) chose to discontinue drug, 4/29 (14%) modified their dose from 1500 to 1000 mg and then discontinued, 9/29 (31%) reduced the dose to 1000 mg and continued and 2/29 (7%) reduced the dose to 500 mg and continued. At the end of the study, of the 31 patients still taking the medication 3 (10%) were taking 500 mg, 9 (29%) were taking 1000 mg, and 19 (61%) were taking 1500 mg. No single reported side effect was correlated with drug termination. However, those who noted either nausea/vomiting/diarrhea were more likely to reduce their dose (4/5 vs. 12/38 p = .04, 9/13 vs. 7/30 p = .004, respectively).

We identified in a real-world cohort of adult patients several barriers to initiation of therapy, confirmed the hematological response, and expanded the range of side effects for voxelotor. Of the 65 patients prescribed voxelotor, 22 never started, and 12 discontinued due to side effects, leaving less than half of the patients prescribed voxelotor taking it a year later, and of those 12/31 were taking reduced doses. We have since developed recommendations to enable more patients to receive medication and to reduce side effect related discontinuation, such as discussing common side effects with each patient before starting the medication, recommending taking the medication with food, using Tylenol or NSAIDs for headaches, and using loperamide and discontinuing stool softeners to treat diarrhea. Many patients noted abdominal complaints resolved entirely after 2 weeks and some were able to increase their dose back to 1500 mg after this.

The most common barrier to starting was that to receive the medication patients must fill out a Patient Authorization and Consent form with GBT. Patients struggled to fill this form out at home, felt uncomfortable signing paperwork, and were reluctant to respond to follow up phone calls from GBT. We began filling out the form in clinic and advised patients to put GBT contact information into their phones so they would be less hesitant to respond to calls from unknown telephone numbers. While we were able to help many patients by coordinating the process in clinic, it must be recognized that filling out all forms with the patients in clinic and addressing insurance refusals and applying for financial assistance represents a significant burden on both patients and providers.

Patients who started voxelotor had a hematologic response similar to that of the HOPE trial despite 14% having received transfusions in the past 3 months and 23% receiving an ESA,¹ both exclusion criteria for enrollment in the Hope trial.

Though all of our VOX+ patients confirmed taking voxelotor in the EMR, only 78% of those with a capillary hemoglobin electrophoresis showed splitting of the A2 or F peaks, or the presence of a shoulder on the S peak characteristic of voxelotor use.² Voxelotor has a short half-life of 35.5 h, so missing one or two doses could be sufficient to lose the effect on capillary electrophoresis.⁵ Also, the studies showing the effects of voxelotor on capillary electrophoresis were in vitro studies; sensitivity and specificity of these studies in the clinical setting is unknown.²

This study has several limitations: it took place at a single SCD center and may not be representative of other regions and care models. The study was retrospective via review of EMR.

This study of voxelotor prescription and use in a cohort of patients with SCD revealed barriers to obtaining medication, changes in laboratory values previously not reported, and high rates of discontinuation and dose modification due to side effects. As voxelotor use becomes more widespread, we can expect to learn more about how to use it to obtain the most benefits for people living with SCD.

致编辑：

2019年，voxelotor被美国食品药品监督管理局批准用于治疗≥16岁患者的镰状细胞病（SCD）。HOPE 试验表明，使用 voxelotor 治疗可减少溶血并改善贫血，不良和严重不良事件的发生率在高剂量组、低剂量组和安慰剂组之间没有差异。^1个然而，临床试验并未提供医生了解该药物在临床实践中的效果所需的所有信息。批准后的真实世界研究可以识别使用障碍以及初始临床研究中未认识到的风险和益处。我们试图检查 voxelotor 在一个综合性城市镰状细胞中心随访的真实世界患者队列中的使用情况。我们假设经济和社会障碍可能会阻止一些患者获得 voxelotor，药物依从性可能不是最佳的，并且不良事件或副作用可能导致临床和血液学益处少于 HOPE 研究中所见。

数据是从我们中心的电子病历 (EMR) 中回顾性获得的。2019 年 11 月 25 日至 2020 年 12 月 31 日期间接受过 voxelotor 处方的 SCD 患者被纳入该分析。羟基脲和红细胞生成刺激剂 (ESA) 的使用由 EMR 中是否存在有效处方来定义。从 EMR 中提取人口统计学、临床结果和 SCD 相关的发病率。

Voxelotor 治疗组包括口头确认获得并开始服药 (VOX+) 的患者，所有其他患者定义为 Vox-。在订购 voxelotor 之前获得实验室测试结果，处于稳定状态，定义为没有急性投诉的访问，并且距离最后一次医院或急诊室 (ED) 使用至少 14 天。对于 VOX+ 患者，在开始使用 voxelotor 后至少 14 天获得了第二组实验室。根据先前研究中描述的方法，通过毛细管区带电泳的变化确定患者血浆中是否存在 voxelotor。^2-4在治疗开始前 3 个月和治疗开始后 3 个月记录导致就诊 ED 或入院的事件以及浓缩红细胞输注的次数。收集了有关副作用、剂量调整、最终停药原因的数据。

使用连续变量的 Mann Whitney U检验和分类变量的 Fisher 精确检验，比较了 VOX+ 和 VOX- 组之间的人口统计学、使用的药物、SCD 相关的发病率、基线急诊就诊、入院、输血数据和基线实验室。使用 Mann Whitney U检验将基线实验室、急诊就诊、入院和输血与 voxelotor 后的实验室、急诊就诊、入院和输血进行比较。使用 Fisher 精确检验检查副作用与停药和剂量调整的相关性。使用 Stata 软件（版本 15）进行分析。

Between November 2019 and December 2020, 65 adults with HbSS/HbSβ0 were prescribed voxelotor; 57% female; mean age 40.7 ± 12.2 years old, 58% on hydroxyurea, 14% had received at least one red blood cell transfusion in the previous 3 months and 19% were on an ESA. Of the 65, 43 (66%) patients started the medication during the time of this study, at a mean of 48.4 days (SD 53.2 days) after it was prescribed (VOX+), while 22 (34%) never started (VOX−).

未开始 (VOX−) 的最常见原因是无法完成该药物的申请流程，因为该药物是在全球血液治疗 (GBT) 护理协调人员的协助下从少数专业药房分发的 (9/22, 41) %)。其他原因包括缺乏保险批准 (4/22, 18%)、高共付额 (1/22, 5%)、失访 (4/22, 18%)、患者拒绝治疗 (1/22, 5 %)、寻求替代疗法的决定 (2/22, 9%) 和患者死亡 (1/22, 5%)。

在 43 名 VOX+ 患者中，39 名以 1500 mg 的剂量开始，3 名以 1000 mg 开始，1 名以 500 mg 开始。VOX− 患者更有可能具有已知的 α 地中海贫血突变，而 VOX+ 患者更有可能有下肢溃疡病史和较高的基线天冬氨酸氨基转移酶。所有其他特征在 VOX+ 和 Vox- 患者之间相似（表 S1）。

血红蛋白的中位数增加为 1.5 g/dL，22/33 (66%) 增加了 1 g/dL 或更多（表 1，图 S1）。实验室数据显示 %HbF 下降 ( p  = .0018)，但总 HbF 没有变化 ( p  = .35)。这表明 F 百分比的变化可能是稀释的，因为血红蛋白 S 由于溶血减少而在循环中存活更长时间。ALT 的统计显着增加 ( p = .0021) 被观察到，这在 Hope 试验中没有被报道过。虽然该队列中没有患者需要为此调整剂量，但最近发现的三名患者在开始用药后 2 周内出现了 2 级 (2) 和 3 级 (1) ALT 异常，所有患者在剂量减少或停药后均有所改善。因此，我们建议监测肝功能测试，并预计在开始使用 voxelotor 后血红蛋白 F 百分比会略有下降。在报告服用 Voxelotor 的患者中，有 78% 的患者在血浆中检测到了 Voxelotor。输血频率 ( p  = .4)、急诊就诊 ( p  = .79) 或入院频率 ( p  = .32)没有变化。

在本次审查时，患者平均接受 voxelotor 治疗 48 周。在这些人中，29/43 (67%) 注意到至少一种副作用，14/43 (33%) 注意到多种副作用。在经历过副作用的人中：8/29 (28%) 选择停药，4/29 (14%) 将剂量从 1500 毫克修改为 1000 毫克然后停药，9/29 (31%) 将剂量减少至 1000 毫克mg 并继续服用，2/29 (7%) 将剂量减少至 500 mg 并继续服用。研究结束时，仍在服用药物的 31 名患者中，3 名 (10%) 服用 500 毫克，9 名 (29%) 服用 1000 毫克，19 名 (61%) 服用 1500 毫克。没有单一报告的副作用与药物终止相关。然而，那些注意到恶心/呕吐/腹泻的人更有可能减少他们的剂量（4/5 对 12/38 p  = .04，9/13 对 7/30 p = .004，分别）。

我们在真实世界的成年患者队列中确定了开始治疗的几个障碍，确认了血液学反应，并扩大了 voxelotor 的副作用范围。在 65 名服用 voxelotor 的患者中，22 名从未开始服用，12 名因副作用而停药，一年后服用 voxelotor 的患者中只有不到一半，其中 12/31 的患者正在减少剂量。从那以后，我们制定了一些建议，以使更多患者能够接受药物治疗并减少与副作用相关的停药，例如在开始用药前与每位患者讨论常见的副作用，建议随餐服药，使用泰诺或非甾体抗炎药治疗头痛，以及使用洛哌丁胺和停用大便软化剂来治疗腹泻。

最常见的启动障碍是接受药物治疗的患者必须填写带有 GBT 的患者授权和同意书。患者在家中难以填写此表格，在文书工作上签字感到不自在，并且不愿回应 GBT 的跟进电话。我们开始在诊所填写表格，并建议患者将 GBT 联系信息存入他们的手机中，这样他们就不会在接听来自未知电话号码的电话时犹豫不决。虽然我们能够通过协调诊所流程来帮助许多患者，但必须认识到，在诊所与患者一起填写所有表格并解决保险拒绝和申请经济援助对患者和提供者来说都是一个沉重的负担。

开始使用 voxelotor 的患者的血液学反应与 HOPE 试验相似，尽管 14% 的患者在过去 3 个月内接受了输血，23% 的患者接受了 ESA，1 这两项都是 Hope 试验的入组排除^标准。

尽管我们所有的 VOX+ 患者都在 EMR 中确认服用了 voxelotor，但只有 78% 的毛细血管血红蛋白电泳显示 A2 或 F 峰分裂，或者在使用 voxelotor 的 S 峰上出现肩峰。² Voxelotor 的半衰期很短，仅为 35.5 小时，因此缺少一剂或两剂就足以失去对毛细管电泳的影响。⁵此外，显示 voxelotor 对毛细管电泳影响的研究是体外研究；这些研究在临床环境中的敏感性和特异性尚不清楚。^2个

这项研究有几个局限性：它发生在一个单一的 SCD 中心，可能不代表其他地区和护理模式。通过审查 EMR，该研究是回顾性的。

这项关于 voxelotor 处方和在一组 SCD 患者中使用的研究揭示了获得药物的障碍、以前未报告的实验室值的变化，以及由于副作用导致的高停药率和剂量调整率。随着 voxelotor 的使用越来越广泛，我们可以期待更多地了解如何使用它来为 SCD 患者获得最大益处。