To the Editor:

More than 40% of women have iron-deficiency anemia (IDA) during pregnancy.¹ Gestational IDA is associated with increased risks of pregnancy complications such as preeclampsia, preterm birth, stillbirth, and maternal mortality.¹ Venous thromboembolism (VTE), including deep vein thrombosis (DVT) and pulmonary embolism (PE), occurs significantly more during pregnancy, with a peak during the postpartum period.² IDA has been shown to be associated with thrombotic events in tumor patients.³ Also, recent animal studies demonstrated an elevated thrombotic risk in iron-deficient mice with thrombus size correlating with platelet count.⁴ During normal pregnancy the hemostatic balance changes in the direction of hypercoagulability to decrease bleeding complications in connection with delivery. In pregnancy, the association between IDA and thrombotic events has not been studied. We studied the association between IDA at admission for delivery and the postpartum VTE risk associated, as well as the potential pathway through thrombocytosis.

We used a large institutional electronic medical records dataset to describe the rate of thrombocytosis in women at admission for delivery and the frequency of postpartum VTE. This dataset included all women delivering at Shanghai First Maternity and Infant Hospital (SFMIH), Tongji University School of Medicine from Jan. 1, 2019 to Dec. 31, 2021. This study was approved by the ethics review board at SFMIH (KS21270). Women who underwent medication thromboprophylaxis (less than 2 weeks before admission) or were diagnosed with DVT or PE before admission were excluded from our analysis. Women with other types of anemia such as thalassemia, infectious and inflammatory anemia, genetic hemoglobin disorders, and other hematological disorders were also excluded. As a routine, women underwent a whole blood analysis at the admission for delivery. Hemoglobin (Hb) levels were prespecified as 10 categories, the highest ≥130 g/L (category 1) and lowest categories ≤89 g/L (category 10), the decrement was 5 g/L between categories. Thrombocytosis was defined as platelet count above the laboratory reference range (>300 × 109/L). The outcome was new-onset VTE including DVT and PE within the first six weeks after delivery. DVT was diagnosed by compression ultrasonography of the proximal veins, while PE was diagnosed based on pulmonary computed tomographic (CT) angiography or clinical diagnosis when the maternal mortality occurred without other explanation (one woman in this study).

All women delivered in this institution had postpartum thrombosis risk assessment scores which were recorded in the electronic medical system. These women who had a score ≥2 and without the contraindications of thromboprophylaxis underwent low molecular weight heparin treatment according to 2015 RCOG's thromboprophylaxis guidelines.⁵ Comprehensive quality control was done by performing a detailed manual chart review of 500 total patients selected at random to validate the recorded thrombosis risk score and thromboprophylaxis. At this check, we found that 16 women (3.2%) had underscored the thrombosis risk and should have undergone low molecular weight heparin thromboprophylaxis.

To analyze the associations of Hb levels with VTE and thrombocytosis. We first described the frequencies and their exponential trends of thrombocytosis and postpartum VTE across the Hb distributions, then logistic regression analysis was used to calculate the adjusted odds ratio (aOR) for postpartum VTE and thrombocytosis associated with categorical Hb levels in two models. We also calculated aORs for postpartum VTE associated with thrombocytosis with or without the adjustment of Hb levels in two models. Model 1 included the individual postpartum thrombosis risk assessment score and low molecular weight heparin thromboprophylaxis. Model 2 included the confounding factors: maternal age, smoking, pre-pregnant body mass index (PPBMI), IVF-ET conception, parity ≥3, multiple pregnancies, preterm, pre-gestational diabetes mellitus, preeclampsia, delivery mode, and postpartum hemorrhage ≥1000 ml or a need of blood transfusion. All data were analyzed with SPSS software (version 22; IBM Corp, Chicago, IL). A probability value of <0.05 was considered statistically significant.

During the study period, there were 79 536 deliveries in our hospital. After restricting the sample to exclusion criteria we had data on 78 180 (98.3%) deliveries in 74 168 unique women. Baseline and obstetric characteristics of the study subjects are shown in Table S1. In our cohort, 107 women were diagnosed postpartum VTE with a rate of 1.37 per 1000 deliveries. Postpartum VTE was more likely to be seen in women delivered by cesarean and preterm, and with multiple pregnancy, preeclampsia, and diabetes. The postpartum hemorrhage rate was also higher in the VTE group. Women in the VTE group had higher postpartum thrombosis risk assessment scores and accordingly more underwent low molecular weight heparin thromboprophylaxis treatment. With respect to hematologic parameters at admission, women who developed VTE had higher platelet counts and slightly lower hemoglobin levels, and accordingly, lower hematocrit, mean corpuscular volume (MCV), and mean corpuscular hemoglobin concentration (MCHC) (Table S2).

The whole blood analysis was performed 1.0 (0.0–1.0) (median and interquartile) days before delivery. The frequency of postpartum VTE across the ten Hb categories is shown in Figure 1. The rate of postpartum VTE increased continuously across the Hb categories when the Hb was <110 g/L (18.7th percentile), with aORs of 1.5 in category 6 (Hb, 105–109 g/L), 1.9 in category 7 (Hb, 100–104 g/L), 3.1 in category 8 (Hb, 95–99 g/L), 4.3 in category 9 (Hb, 90–94 g/L), and 6.2 in category 10 (Hb ≤89 g/L), respectively. The results were similar in the two adjusted models. (Table S3). The overall rate of thrombocytosis was 7.4%. The frequency of thrombocytosis increased continuously, particularly when the Hb was <110 g/L, across the ten Hb categories (Figure 1) (Table S3). Compared to category 1, the aOR of thrombocytosis associated with category 10 was 16.5. Thrombocytosis was associated with increased risk of postpartum VTE with aOR of 1.8 without adjustment of Hb levels, nevertheless, this association was attenuated with the additional adjustment of Hb levels. The results were similar in the two adjusted models (Table S4).

Our study yielded two key findings: (1) The rate of postpartum VTE increased continuously across the Hb categories when the Hb was <110 g/L. (2) The frequency of thrombocytosis increased continuously with decreasing Hb levels in pregnant women.

Anemia in pregnancy is a common and important public health concern around the world. Iron deficiency is the most common cause of anemia owing to the increased iron demand and physiological hemodilutional changes during pregnancy. IDA has been shown to be associated with VTE in certain populations such as patients with cancer. The thrombotic risk was increased in iron-deficiency mice where thrombus size was positively related to platelet count, suggesting the elevated thrombotic risk in IDA could be partially attributed to reactive thrombocytosis in animals.⁴ In the limited human analyses, Song et al observed a 2-fold thrombosis risk in patients with IDA and thrombocytosis compared with patients with IDA alone in a large scale of elder population.⁶ Our non-linear analysis indicated strong, continuous associations of Hb levels below 110 g/L with increased risks of thrombocytosis and postpartum VTE in a similar trend. In our analysis, thrombocytosis was associated with an increased risk of postpartum VTE without adjustment of Hb levels, nevertheless, this association became non-significant when considering Hb levels in the model, which is likely to be explained by the correlation between IDA and thrombocytosis. Our data suggest that reactive thrombocytosis itself is in the pathway of IDA to increase the risk of VTE in pregnancy.

We first assessed the association of Hb levels in pregnancy and postpartum VTE in large-volume clinical data. In order to control the confounding bias, we used two models to adjust the risks of VTE and thrombocytosis associated with categorical Hb and the results were similar in these two models. The main limitation of our study is that we supposed that low Hb concentrations were caused by iron deficiency due to the lack of serum ferritin concentration. Another limitation is that we could not know whether iron supplements during pregnancy could reverse the risks of postpartum VTE associated with low Hb or not, which needs further study in the future. Recognition of this increased thrombotic risk of IDA in pregnancy raises the question of adding this factor to the existing additive ordinal point based on the scoring system for potential postpartum thromboprophylaxis. We did not recommend this possibility for several reasons. The IDA is very common during pregnancy, and the number needed to treat to prevent one VTE is likely very high given the modestly increased individual thrombotic risk. The IDA itself is just easy to correct and should be the focus of treatment. Recognition and prompt treatment of IDA is more important during pregnancy, not only because of the potential increased thrombotic risk, but also other maternal and neonatal morbidity such as preeclampsia, preterm and postpartum hemorrhage, which were also demonstrated in our study.

In conclusion, we observed IDA during pregnancy was associated with increased risk of postpartum VTE. Given the global burden of unrecognized and undertreated IDA during pregnancy, our findings emphasize the clinical implication of prompting identification and appropriate treatment of this common hematologic disorder.

致编辑：

超过 40% 的女性在怀孕期间患有缺铁性贫血 (IDA)。¹妊娠 IDA 与先兆子痫、早产、死产和孕产妇死亡率等妊娠并发症的风险增加有关。¹静脉血栓栓塞症 (VTE)，包括深静脉血栓形成 (DVT) 和肺栓塞 (PE)，在怀孕期间发生率明显更高，并在产后期间达到高峰。² IDA 已被证明与肿瘤患者的血栓事件有关。³此外，最近的动物研究表明，缺铁小鼠的血栓形成风险升高，血栓大小与血小板计数相关。^4个在正常妊娠期间，止血平衡向高凝状态方向改变，以减少与分娩相关的出血并发症。在怀孕期间，尚未研究 IDA 与血栓事件之间的关联。我们研究了分娩时 IDA 与产后 VTE 风险之间的关系，以及通过血小板增多的潜在途径。

我们使用了一个大型机构电子病历数据集来描述女性入院时的血小板增多率和产后静脉血栓栓塞症的发生率。该数据集包括 2019 年 1 月 1 日至 2021 年 12 月 31 日在同济大学医学院附属上海市第一妇婴医院 (SFMIH) 分娩的所有女性。本研究已获得 SFMIH 伦理审查委员会的批准 (KS21270)。接受药物血栓预防（入院前不到 2 周）或入院前被诊断患有 DVT 或 PE 的女性被排除在我们的分析之外。患有其他类型贫血如地中海贫血、感染性和炎症性贫血、遗传性血红蛋白病和其他血液病的女性也被排除在外。作为一项常规工作，女性在入院分娩时接受了全血分析。血红蛋白 (Hb) 水平预先指定为 10 类，最高≥130 g/L（第 1 类）和最低类别≤89 g/L（第 10 类），类别之间递减 5 g/L。血小板增多症定义为血小板计数高于实验室参考范围 (>300 × 109/L)。结果是新发 VTE，包括分娩后前六周内的 DVT 和 PE。DVT 是通过近端静脉加压超声诊断的，而 PE 是根据肺部计算机断层扫描 (CT) 血管造影或临床诊断，当产妇死亡而没有其他解释时（本研究中有一名女性）。血小板增多症定义为血小板计数高于实验室参考范围 (>300 × 109/L)。结果是新发 VTE，包括分娩后前六周内的 DVT 和 PE。DVT 是通过近端静脉加压超声诊断的，而 PE 是根据肺部计算机断层扫描 (CT) 血管造影或临床诊断，当产妇死亡而没有其他解释时（本研究中有一名女性）。血小板增多症定义为血小板计数高于实验室参考范围 (>300 × 109/L)。结果是新发 VTE，包括分娩后前六周内的 DVT 和 PE。DVT 是通过近端静脉加压超声诊断的，而 PE 是根据肺部计算机断层扫描 (CT) 血管造影或临床诊断，当产妇死亡而没有其他解释时（本研究中有一名女性）。

所有在该机构分娩的妇女都有产后血栓形成风险评估评分，这些评分记录在电子医疗系统中。根据 2015 RCOG 的血栓预防指南，这些评分≥2 且没有血栓预防禁忌症的女性接受了低分子肝素治疗。⁵全面的质量控制是通过对随机选择的 500 名患者进行详细的人工图表审查来完成的，以验证记录的血栓形成风险评分和血栓预防。在这次检查中，我们发现 16 名女性 (3.2%) 强调了血栓形成的风险，应该接受低分子肝素血栓预防。

分析 Hb 水平与 VTE 和血小板增多症的关系。我们首先描述了 Hb 分布中血小板增多症和产后 VTE 的频率及其指数趋势，然后使用逻辑回归分析计算了两个模型中与分类 Hb 水平相关的产后 VTE 和血小板增多症的调整优势比 (aOR)。我们还计算了与血小板增多症相关的产后 VTE 的 aOR，在两个模型中调整或不调整 Hb 水平。模型 1 包括个体产后血栓形成风险评估评分和低分子肝素血栓预防。模型 2 包括混杂因素：母亲年龄、吸烟、孕前体重指数 (PPBMI)、IVF-ET 受孕、胎次≥3、多胎妊娠、早产、妊娠前糖尿病、先兆子痫、分娩方式，产后出血≥1000毫升或需要输血。所有数据均使用 SPSS 软件（第 22 版；IBM Corp, Chicago, IL）进行分析。<0.05 的概率值被认为具有统计学意义。

研究期间，我院共分娩 79 536 例。在将样本限制为排除标准后，我们​​获得了 74 168 名独特女性的 78 180 次（98.3%）分娩数据。研究对象的基线和产科特征如表 S1 所示。在我们的队列中，107 名妇女被诊断为产后 VTE，每 1000 次分娩中有 1.37 例发生。产后 VTE 更可能发生在剖宫产和早产、多胎妊娠、先兆子痫和糖尿病的妇女中。VTE 组的产后出血率也更高。VTE 组的女性产后血栓形成风险评估评分较高，因此更多接受低分子肝素血栓预防治疗。关于入院时的血液学参数，

全血分析在分娩前 1.0 (0.0–1.0)（中位数和四分位数）天进行。十个 Hb 类别的产后 VTE 频率如图 1 所示。当 Hb <110 g/L（第 18.7 个百分位）时，产后 VTE 的发生率在 Hb 类别中持续增加，类别 6 中的 aOR 为 1.5（ Hb, 105–109 g/L), 1.9 在类别 7 (Hb, 100–104 g/L), 3.1 在类别 8 (Hb, 95–99 g/L), 4.3 在类别 9 (Hb, 90–94 g/L）和类别 10（Hb ≤89 g/L）中的 6.2。两个调整模型的结果相似。（表 S3）。总的血小板增多率为7.4%。在十个 Hb 类别中，血小板增多的频率持续增加，尤其是当 Hb <110 g/L 时（图 1）（表 S3）。与类别 1 相比，与类别 10 相关的血小板增多症的 aOR 为 16.5。在不调整 Hb 水平的情况下，血小板增多与产后 VTE 风险增加相关，aOR 为 1.8，然而，这种关联随着 Hb 水平的额外调整而减弱。两个调整模型的结果相似（表 S4）。

我们的研究得出了两个关键发现：(1) 当 Hb <110 g/L 时，Hb 类别的产后 VTE 发生率持续增加。(2)随着孕妇Hb水平的降低，血小板增多的频率不断增加。

妊娠期贫血是全世界普遍且重要的公共卫生问题。由于怀孕期间铁需求增加和生理性血液稀释变化，缺铁是贫血的最常见原因。IDA 已被证明与某些人群（如癌症患者）的 VTE 相关。缺铁小鼠的血栓形成风险增加，其中血栓大小与血小板计数呈正相关，这表明 IDA 中血栓形成风险升高可部分归因于动物的反应性血小板增多症。⁴在有限的人体分析中，Song 等人在大规模老年人群中观察到 IDA 和血小板增多症患者的血栓形成风险是单纯 IDA 患者的 2 倍。^6个我们的非线性分析表明，低于 110 g/L 的 Hb 水平与血小板增多症和产后静脉血栓栓塞症的风险增加有很强的持续关联，趋势相似。在我们的分析中，在不调整 Hb 水平的情况下，血小板增多与产后 VTE 风险增加相关，然而，当考虑模型中的 Hb 水平时，这种关联变得不显着，这可能是由 IDA 和血小板增多之间的相关性解释的。我们的数据表明，反应性血小板增多症本身是 IDA 增加妊娠 VTE 风险的途径。

我们首先在大量临床数据中评估了妊娠期间 Hb 水平与产后 VTE 的关联。为了控制混杂偏倚，我们使用两种模型来调整与分类 Hb 相关的 VTE 和血小板增多症的风险，这两种模型的结果相似。我们研究的主要局限性在于我们假设低 Hb 浓度是由于血清铁蛋白浓度不足导致缺铁引起的。另一个限制是我们无法知道怀孕期间的铁补充剂是否可以逆转与低 Hb 相关的产后 VTE 风险，这需要在未来进一步研究。对妊娠期 IDA 血栓形成风险增加的认识提出了将这一因素添加到基于潜在产后血栓预防评分系统的现有附加序数点的问题。出于多种原因，我们不推荐这种可能性。IDA 在怀孕期间非常常见，鉴于个体血栓形成风险适度增加，预防一次 VTE 所需治疗的人数可能非常高。IDA本身只是容易纠正，应该是治疗的重点。IDA 的识别和及时治疗在怀孕期间更为重要，这不仅是因为潜在的血栓形成风险增加，而且还因为其他孕产妇和新生儿的发病率，如先兆子痫、早产和产后出血，我们的研究也证明了这一点。出于多种原因，我们不推荐这种可能性。IDA 在怀孕期间非常常见，鉴于个体血栓形成风险适度增加，预防一次 VTE 所需治疗的人数可能非常高。IDA本身只是容易纠正，应该是治疗的重点。IDA 的识别和及时治疗在怀孕期间更为重要，这不仅是因为潜在的血栓形成风险增加，而且还因为其他孕产妇和新生儿的发病率，如先兆子痫、早产和产后出血，我们的研究也证明了这一点。出于多种原因，我们不推荐这种可能性。IDA 在怀孕期间非常常见，鉴于个体血栓形成风险适度增加，预防一次 VTE 所需治疗的人数可能非常高。IDA本身只是容易纠正，应该是治疗的重点。IDA 的识别和及时治疗在怀孕期间更为重要，这不仅是因为潜在的血栓形成风险增加，而且还因为其他孕产妇和新生儿的发病率，如先兆子痫、早产和产后出血，我们的研究也证明了这一点。

总之，我们观察到妊娠期间的 IDA 与产后 VTE 风险增加有关。鉴于妊娠期间未被识别和治疗不足的 IDA 的全球负担，我们的研究结果强调了促进识别和适当治疗这种常见血液病的临床意义。