下腔静脉滤器回收失败的预测因素分析

From the Society for Clinical Vascular Surgery

Factors predicting failure of retrieval of inferior vena cava filters

--来自美国临床血管外科学会

下腔静脉滤器回收失败的预测因素分析

Katherine L. Morrow. MS. James Bena, Ms, Sean P. Lyden. MD, Ezequiel Paradi, MD, and Christopher J. Smolock, MD, Cleveland. Ohiof

ABSTRACT 摘要

Objective: Inferior vena cava (IVC) filters have been commonly used to prevent pulmonary embolism in patients with deep vein thrombosis. However, IVC filters have been associated with risks, including IVC perforation, filter migration, fracture, and thrombosis. Filter retrieval has not always been successful. Our objective was to identify the factors associated with failure of retrieval of IVC filters.

目的:下腔静脉滤器是预防深静脉血栓患者肺栓塞的常用方法。然而腔静脉滤器置入存在下腔静脉穿孔、滤器移位、滤器折断及滤器内血栓形成等风险。不是所有置入体内的滤器都能成功取出的，本研究的目的是找出与滤器回收失败相关的影响因素。

Methods: The present study was an institutional review board-approved retrospective medical record review of patients who had undergone IVC filter retrieval attempts at the Department of Vascular Surgery at Cleveland Clinic from 2011 to 2018. The patients were identified by International Classification of Diseases code query, and data were gathered regarding demographics, filter position, procedure details, and patient outcomes. Computed tomography imaging and venography was used to determine the IVC filter location before retrieval.

方法:本研究是一项伦理委员会批准的回顾性研究，通过就诊医疗文书分析了2011年至2018年在克利夫兰诊所血管外科（尝试）进行腔静脉滤器回收的所有患者资料。采用国际疾病分类编码筛选患者，主要搜集滤器形态、位置、操作及患者预后等数据。滤器回收前行CT和静脉造影检查以确定腔静脉滤器的位置。

Results: We identified 295 filter retrieval attempts in 294 patients. No procedural IVC ruptures, morbidity, or mortality occurred. Retrieval was successful for 249 filters (84.4%). The median filter dwell time was 196 days for the successful retrievals compared with 375 days for the failed retrieval attempts (P=.004). Penetration of the filter tines through the caval wall occurred in 291 filters (98.6%). However, the hook/apex (HA) of 31 filters (10.5%) had become embedded or had penetrated through the caval wall. The hook/apex and collar (HA+C) of 33 filters (11.2%) were embedded or had penetrated through the caval wall. The failure rate of filter retrieval with the HA embedded was 48.4% (15 of 31). The failure rate with the HA+C embedded was 66.7% (22 of 33). Finally, the failure rate for filters without these issues was 3.9% (9 of 231). The failure rate for HA and HA+C was greater than that for those without these issues (P < .001) but did not differ between the two issues (P =.14). Among those with computed tomography scans, the association of any portion of the filter with other adjacent retroperitoneal structures was not related to an increased rate of retrieval failure (P=.16). Complex retrieval methods involving endobronchial forceps, ballooning, or snaring the collar of the filter was associated with increased retrieval failure compared with simple retrieval involving snaring the hook of the filter (P < .001). The failure rates decreased over time (P=.004). Of the patients with failed retrieval attempts, 8% experienced subsequent venous thromboembolism.

结果:共294例患者接受了295次滤器（尝试）取出术，无操作相关的滤器断裂、医源性损伤或患者死亡，249个滤器被成功回收(84.4%)。成功取出的滤器体内停留平均时间为196天，在研究时段内未能取出者平均置入时间为375天(P=0.004)。发现291个滤器(98.6%)的滤齿穿透腔静脉壁，但是只有31个滤器(0.5%)的回收钩/顶端(HA)嵌入或穿透下腔壁。33个滤器(11.29%)的回收钩/顶点和领口部(HA+C)嵌入或穿透管腔壁。回收钩/顶端(HA)嵌入的滤器回收失败率为48.4%(5 / 31)。回收钩/顶端及领口部同时嵌入血管壁（HA+C）的失败率为66.7%(33例中有22例)。最后，没有这些问题的滤器回收失败率是3.9%(9/231)。存在HA和HA+C嵌入的回收失败率比没有这些问题的过滤器的失败率大(P< .001)，但HA和HA+C之间没有统计学差异(P=.14)。CT扫描发现滤器的任一部分与其邻近的腹膜后结构的相对位置关系与回收失败率增加无关(P=.16)。与标准的圈套滤器取出钩相比应用活检钳、球囊剥离及圈套器圈套滤器衣领部等方法的应用与高更高的回收失败率正相关 (P < 0.01)。回收失败率随着时间推移会下降(P ＜ 0.04)，8%滤器回收失败的患者后续并发了血栓栓塞事件。

Conclusions: Retrieval should be attempted for all IVC filters, irrespective of the chronicity and complexity, given the procedural safety. Tine penetration was nearly ubiquitous; however, IVC filters with the HA or HA+C imbedded into or penetrating through the caval wall was a predictor of retrieval failure.

结论:不考虑滤器留置的长期性和复杂性的话，所有腔静脉滤器都应该回收。滤器齿穿透血管壁普遍存在，滤器回收钩嵌入（穿透）血管壁（HA）或滤器回收钩和滤器领口部（HA+C）同时嵌入或穿透腔静脉壁是回收失败的预测因子。

Keywords: Deep vein thrombosis; Vena cava filter; Hook/apex and collar; Inferior vena cava; Pulmonary embolism

【关键词】深静脉血栓形成；腔静脉滤器；回收钩/滤器领口部；下腔静脉；肺栓塞

Inferior vena cava (IVC) filters have been commonly used to prevent pulmonary embolism (PE) in patients with deep vein thrombosis (DVT), despite inconclusive evidence regarding the overall effects on morbidity and mortality posed by these devices. Although IVC filters have been shown to decrease the risk of PE in patients with a contraindication to anticoagulation, filters have also been associated with a range of risks, including IVC perforation, filter migration, fracture, and thrombosis. At present, no clinical practice guidelines are available regarding the optimal timeline for filter retrieval. The rates of filter retrieval attempts have varied widely in the reported studies, and when retrieval has been attempted has not always been successful. The retrieval attempt rates appear to have been increasing modestly over time but have remained lower than the ideal. Although filter retrieval failure has been associated with an increased filter dwell time, the effect of the filter position relative to the caval wall and surrounding structures has not been fully elucidated. Thus, the objective of the present study was to identify the factors associated with failure to retrieve IVC filters.

下腔静脉(IVC)滤器通常用于预防深静脉血栓(DVT)患者的肺栓塞(PE)，虽然关于这些医疗器具对患者发病率和死亡率的总体影响尚无确凿证据。腔静脉滤器已被证明可降低抗凝禁忌患者肺栓塞的风险，但腔静脉滤器也与一系列风险相关，主要包括静脉血管穿透、滤器移位、折断和滤器内血栓形成。目前，还没有临床指南界定最佳滤器回收时机。在已发表的研究中，滤器回收率各不相同，并且滤器不是每个都能成功取出的。随着时间推移尝试取出率会稍有增加，但整体取出率仍不理想。滤器滞留体内时间随着回收失败而显著增加，滤器的位置对腔静脉壁和周围临近组织影响尚未完全明了，本研究旨在阐述滤器回收失败的影响因素。

METHODS 方法

The present study was a retrospective medical record review of all patients who had undergone IVC filter retrieval attempts at the Department of Vascular Surgery at Cleveland Clinic from July 2011 through June 2018. The institutional review board approved the present study and allowed for a waiver of informed consent owing to the minimal risk posed to the patients. The patients were identified by International Classification of Disease code query, and patient data were gathered from the electronic medical records. The data collected from the patients’ medical records included demographics, medical history, indication for filter placement, filter type, history of filter thrombosis, institution of filter placement (Cleveland Clinic or outside hospital), filter dwell time, American Society of Anesthesiologists class at the retrieval attempt, and retrieval method (eg, snaring the filter hook or more complex methods such as endobronchial forceps or ballooning).

回顾性分析2011年7月至2018年6月在克利夫兰诊所血管外科接受下腔静脉滤器取出尝试的所有患者病历。机构审查委员会批准了本研究，并允许放弃知情同意，因这对患者构成的风险极小。通过国际疾病分类码查询对患者进行识别，并从电子病历中采集患者数据。从患者病历中收集的数据包括人口统计、病史、滤器放置的适应症、类型、血栓形成史、放置机构（克利夫兰诊所或院外）、滤器体内停留时间、美国麻醉师学会分类检索取出尝试，以及取回方法（例如，勾住回收钩或更复杂的方法，如支气管内镜活检钳或球囊扩张使之与血管壁剥离）。

Computed tomography (CT) and venography were used to determine the IVC filter location before the retrieval attempt. The filter hook/apex (HA) and collar (HA+C) were determined to be intraluminal, within the caval wall, or penetrating through the caval wall (Fig 1).The filter tines were also assessed to determine whether they were intraluminal or whether any of the tines of a given filter had penetrated through the caval wall. Filter association with adjacent structures such as bowel or the spinal column was determined (Fig 2). Additionally, patency of the IVC was categorized as either<50%>50% stenosed,oroccluded. Imaging studies after filter retrieval were reviewed to determine whether any portion of the filter had been left behind in the patient. Simple filter retrieval was defined as snaring the hook of the filter. In contrast, complex retrieval was defined as any combination of the following: snaring the collar of the filter, using a balloon to dislodge the filter, and/or using endobronchial forceps to retrieve the filter (Fig 3).

尝试取出滤器之前完善CT和静脉造影检查以确定滤器的位置。回收钩/顶端（HA）和领口部（HA+C）被判断在腔内、管壁内还是已经穿透管壁（图1）。滤器齿也被评估以确定它们是否为腔内或任一尖端是否穿透腔壁。确定滤器与邻近结构（如肠或脊柱）的关系（图2）。此外，下腔静脉通畅可分为<50%狭窄，>50%狭窄或完全梗阻。检查滤器取出后的图像，以确定滤器断裂部分是否遗留在患者体内。简单的滤器回收定义为圈套器圈套住滤器的回收钩进行回收，而复杂滤器回收定义为任何联合使用了圈套器圈套滤器领口部回收、球囊扩张使之与血管壁剥离后回收或使用内镜活检钳回收（图3）。

Continuous measures are summarized using the mean ± standard deviation or median and quartiles, according to whether the measures were normally distributed using the Shapiro-Wilk test. Categorical factors are described using frequencies and percentages. Comparisons of the risk factors for retrieval success were evaluated using the Pearson X² test for categorical factors and the Kruskal-Wallis test or analysis of variance models for non-normally distributed and normally distributed continuous measures, respectively. Odds ratios for retrieval success with 95% confidence intervals are also presented for each risk factor. Given that only one patient had had multiple filters, no correction for correlation of the responses within the patients was used. Statistical analysis was performed using SAS software, version 9.4 (SAS Institute, Cary, NC). A significance level of P < .05 was assumed for all tests.

使用标准差或中位数和四分位数描述连续测量值，用Shapiro-Wilk检验测量值是否正态分布，分类变量用频率和百分比来描述。用Pearson-X²检验滤器回收成功的危险因素的分类变量，用Kruskal-Wallis检验非正态分布和正态分布回收成功的危险因素的连续变量。对于每个风险因素，也给出了95%置信区间的回收成功的优势比。考虑到只有一名患者有多个滤器，没有对患者间不同反应进行比较。使用SAS软件9.4版（SAS Institute，Cary，NC）进行统计分析。所有试验的显著性水平均假定为P<.05。

ARTICLE HIGHLIGHTS

l Type of Research: Single-center, retrospective cohort study

l Key Findings: The retrieval failure rate of 295 inferior vena cava filters in 294 patients for filters with the hook/apex penetrating through the caval wall was 48.4% (15 of 31), with hook/apex and collar penetration through the caval wall was 66.7% (22 of 33), and without these issues was 3.9% (9 of 231).

l Take Home Message: Inferior vena cava filters with the hook/apex or hook/apex and collar penetrating through the caval wall had increased rates of retrieval failure. Filter tine penetration into adjacent structures did not significantly affect the retrieval failure rates.

文章重点介绍

研究类型：单中心回顾性队列研究

关键发现：294例患者中295个下腔静脉滤器取出钩/尖端穿透下腔静脉壁的回收失败率为48.4%（15/31），回收钩/领口部穿透下腔静脉壁的回收失败率为66.7%（22/33），没有这些问题的是3.9%（231例中的9例）。

重要结论：下腔静脉滤器的钩/尖（HA）或钩/尖和衣领部（HA+C）穿过腔静脉壁增加了回收失败率，滤器齿穿透邻近结构对回收失败率没有显著影响。

图1. 滤器相对于血管壁的位置。A，滤器取出钩和衣领部进入血管壁，滤器齿穿透腔静脉壁。B，回收钩，领口部和滤齿穿透腔静脉管壁

图2。滤器齿穿透腔静脉壁进入邻近结构：A，肝脏；B，肠管；C，主动脉；D，脊柱。

图3. 下腔静脉滤器取出技术：A、圈套住滤器回收钩；B、用球囊将滤器从腔壁分离；C、用支气管内镜活检钳取出滤器。

RESULTS 结果

A total of 295 filter retrieval attempts in 294 patients were identified. The patient demographic data and medical comorbidities are summarized in Tables I and II. Female gender was associated with increased retrieval failure (P=.049); however, race (P=.48) and age (P=.62) were not. The only medical comorbidity we found associated with an increased risk of retrieval failure was hypercoagulability (P=.024).

294例患者共有295次滤器取出尝试。表1和表2总结了患者的人口统计数据和医疗合并疾病数据。女性性别与取出失败率增加有相关性（P=.049）；但是种族（P=.48）和年龄（P=.62）与此无关。我们发现的唯一与取出失败风险增加相关的医学合并症是高凝状态（P=.024）。

The vast majority of filters were temporary (n=282; 95.6%), with the remainder permanent (Table III). Retrieval was successful for 249 of the filters (84.4%). Retrieval success was greater for the temporary filters (85.5%) than for the permanent filters (61.5%; P=.020). Of the temporary filters, retrieval success was greater for the Celect filters (93.5%) and Eclipse/Denali filters (93.8%) than for the Option filters (78.4%; P=.002). No procedural IVC ruptures, morbidity, or mortality occurred during the retrieval attempts or within the first 30 days after the procedure. Ten of the filters (3.5%) had a history of filter thrombosis; however, this was not associated with increased retrieval failure (P=.70).

绝大多数滤器是临时性的（n=282；95.6%），其余的是永久性的（表III）。249个滤器（84.4%）回收成功。临时滤器的回收成功率（85.5%）高于永久滤器（61.5%；P＜0.020）。在临时滤器中，Celect滤器（93.5%）和Eclipse/Denali滤器（93.8%）的回收成功率高于Option滤器（78.4%；P=0.002）。在围术期及手术后的30天内，没有发生操作相关性腔静脉破裂、操作相关病死事件。其中10个滤器（3.5%）有过血栓形成史，然而这与回收失败率增加无关（P=0.70）。

Most of the filters had been placed at our institution (n=227; 76.9%), with the remainder placed at an outside hospital (n=68; 23.1%). The indication for filter placement included a contraindication to anticoagulation (n=100; 34%), bridging for surgery (n=76; 26%), bleeding during anticoagulation therapy (n=41; 24%), clot propagation despite anticoagulation therapy (n=12; 4.1%), prophylaxis for DVT thrombectomy or thrombolysis (n=8; 2.7%), prophylaxis after trauma (n=5; 1.7%), free-floating thrombus in the IVC, iliac vein, or femoral vein (n=4; 1.4%), PE with a limited cardiac reserve (n=3; 1.0%), anticoagulation noncompliance (n=2; 0.7%), and placement despite the lack of a clear indication (n=10; 3.4%). The indication for placement could not be obtained for a subset of the patients whose filter had been placed at an outside hospital (n=34; 12%).

大部分滤器在我们中心放置的（n=227；76.9%），其余的在外部医院放置的（n=68；23.1%）。滤器放置的适应症包括抗凝禁忌症（n= 100；34%）、等待手术（n=76；26%）、抗凝治疗期间出血（n=4 41；24%）、抗凝治疗后血栓仍进展（n =12；4.1%），预防血栓清除或导管溶栓后复发（n＝8；2.7%），创伤后预防血栓（n＝5；1.7%），腔静脉漂浮血栓、髂静脉或股静脉（n=4；1.4%），肺栓塞并且心功能差（n=3；1%），抗凝不依从（n=2；0.7%）；无明确适应症（n=10；3.4%）。滤器于院外放置且无法获得放置指征的（n=34；12%）。

表1. 滤器回收成功的人口统计学预测因素

The median filter dwell time was 196 days for the filters successfully retrieved and 375 days for the filters with failed retrieval attempts (P=.004; Fig 4). The median dwell time for the permanent filters was 3605 days compared with 196 days for the temporary filters (P < .001). Filters with both hook and collar penetrating through the caval wall had had statistically significantly longer dwell times than filters with the intraluminal hook and collar and filters with only the hook penetrating through the caval wall (P < .001). Filters with an intraluminal hook and collar through the caval wall had an average dwell time of 231 days. Filters with the hook through the caval wall had an average dwell time of 287 days. Filters with both the hook and the collar through the caval wall had an average dwell time of 591 days.

对于成功取出的滤器，中位滤器体内停留时间为196天，对于取出尝试失败的滤器，中位滤器体内停留时间为375天（p=.004; Fig 4）。永久滤器的中位体内停留时间为3605天，而临时滤器的中位体内停留时间为196天（P<.001）。滤器取出钩和衣领部穿透管腔壁的体内停留时间明显长于回收钩和衣领部仅进入腔静脉壁的滤器和仅回收钩穿透管腔壁的滤器（P<.001）。回收钩和衣领部进入管腔壁的滤器平均体内停留时间为231天，回收钩穿透管腔壁的滤器平均体内停留时间为287天，回收钩和衣领部同时穿透管腔壁的滤器平均体内停留时间为591天。

表II.滤器回收成功的病史预测因素

Penetration of the filter tines through the caval wall occurred in 291 of the filters (98.6%; Table IV). The HA of 31 filters (10.5%) were embedded or penetrating through the caval wall, and the HA+C of 33 (11.2%) were embedded or had penetrated through the caval wall. The failure rate for the retrieval of filters with the HA embedded was 48.4% (15 of 31), that with the HA+C embedded was 66.7% (22 of 33), and that for filters without these issues was 3.9% (9 of 231). The failure rate of both filters with the HA embedded and those with the HA+C embedded was greater than that of those without these issues (P < .001). However, the failure rates between the two did not differ from one another (P=.14; Fig 5). We found a total of six filters in the present study that had fractured and had left ≥1 tine behind after retrieval. Of these six filters, three had been permanent and three temporary. Neither the size of the retrieval sheath (P=.45) nor the use of the coaxial sheaths (P=.65) for filter removal was associated with increased retrieval failure.

291个滤器（98.6%；表IV）的滤齿穿过腔壁。31个滤过器（10.5%）的HA嵌入或穿透腔壁，33个（11.2%）的HA+C嵌入或穿透腔壁。HA嵌入管腔壁回收失败率为48.4%（15/31），HA+C嵌入失败率为66.7%（22/33），无以上问题者为3.9%（9/231）。HA和HA+C嵌入的滤器的回收失败率均高于无此问题者（P<0.001）。然而，两者之间（HA VS HA+C）的回收失败率并没有差异（P=.14；图5）。在本研究中，我们一共发现了6个滤器断裂，在滤器取出后至少有1个残余结构遗留体内。在这6个滤器中，3个是永久性的，3个是临时性的。回收鞘的大小（p=.45）和是否使用同轴系统（P=. 65）和回收失败率增加无关。

表III. 回收成功的滤器相关预测因素

CT scans of the abdomen before filter retrieval was available for 157 patients (47%). Of these patients, filter tines or hooks were associated with the following structures: bowel (n=50; 32%), spine (n=17; 11%), aorta (n=11; 7%), psoas major muscle (n=6; 3.8%), renal vein (n=5; 3.2%), liver (n=4; 2.5%), heart (n=3; 1.9%), lung (n=2; 1.3%), and pulmonary artery (n=1; 0.6%). Among those with CT scans available, the association of any portion of the filter with other adjacent structures was not related to an increased rate of failed retrieval (P=.16; Table IV). However, six patients had had ≥1 filter tine left behind after retrieval because of filter fracture that had occurred before retrieval. Complex retrieval methods involving endobronchial forceps, ballooning, or snaring the collar of the filter was associated with increased failure rates compared with simple retrieval involving snaring the hook of the filter (P < .001).

157例患者（47%）在滤器取出前进行了腹部CT扫描。在这些患者中，滤齿或回收钩与以下结构相关：肠（n=50；32%）、脊柱（n=17；11%）、主动脉（n=11；7%）、腰大肌（n=6；3.8%）、肾静脉（n=5；3.2%）、肝（n=4；2.5%）、心脏（n=3；1.9%）、肺（n=2；1.3%）和肺动脉（n=1；0.6%）。在那些有CT扫描的患者中，滤器的任何部分与其他邻近结构的关联性与取出失败率的增加无关（P=.16；表IV）。然而，6个病人体内残留至少1个断裂的滤器齿，因为取出前就发现滤器断裂。与简单的圈套滤器回收钩相比，复杂的取回方法包括支气管内镜活检钳，球囊扩张使之与血管壁剥离，或圈套滤器的领口部，这些方法的使用与回收失败率增加正相关（P<.001）。

图4. 尝试回收前滤器体内停留时间

Before filter retrieval, most patients had had a relatively patent IVC. IVC stenosis of<50% 274="" was="" present="" in="" patients="" ivc="" stenosis="" of="">50% was present in 19 patients (6%), and IVC occlusion was seen in 1 patient (0.3%). Stenosis of >50% or complete occlusion was associated with increased failure of retrieval compared with stenosis of<50% (P=.013; Table IV).

在滤器取出之前，大多数患者有一个相对通畅的下腔静脉。274例（93%）下腔静脉狭窄<50%，19例（6%）下腔静脉狭窄>50%，1例（0.3%）下腔静脉闭塞。大于50%的狭窄或完全闭塞与小于50%的狭窄相比，滤器回收失败率增加（P=.013；表IV）。

表IV：第一次尝试取出成功的预测因素

The retrieval failure rates had decreased over time (P=.004), suggesting a learning curve was involved with the retrieval process (Fig 6). Of the patients with failed retrieval attempts, 8% had experienced subsequent venous thromboembolism within the first 2 years after the retrieval attempt. Of these patients, two had experienced PE, but that was unrelated to the filter retrieval attempt or the continued presence of the original DVT. Of the 46 filters that were not successfully retrieved, a second retrieval was later attempted for eight (17%), six of which were successful. Any second attempts at filter retrieval were not included within our data analyses in the present study. One patient within our cohort had required open surgical removal of the IVC filter. This patient’s filter had formed a fistula into the duodenum, resulting in anorexia and postprandial pain that necessitated filter retrieval.

随着时间的推移回收失败率下降了（P＜.004），表明学习曲线与回收过程有关（图6）。在尝试回收失败的患者中，8%的患者在尝试取出后的前2年内发生过静脉血栓栓塞时间。在这些病人中，有两个有肺栓塞，但这与滤器取出尝试或原深静脉血栓的持续存在无关。在未成功取出的46个滤器中，有8个（17%）稍后尝试再次取出，其中6个已成功。在本研究中，我们的数据分析中没有包含任何第二次滤器取出尝试。在我们的队列中有一个病人要求开放手术取出滤器，这个病人的滤器已经形成了一个到十二指肠的瘘，导致厌食和餐后疼痛，需要开放手术将其取出。

DISCUSSION 讨论

IVC filters have long been a method of preventing PE in patients with a contraindication to anticoagulation or other factors that confer a high risk of venous thromboembolism. In general, filters have been considered a relatively safe device compared with the potentially devastating morbidity and mortality associated with PE. However, filters are not without risk, especially because of the tendency to leave filters in place long past the point at which they should have been retrieved. A systematic review of 37 studies and 6834 retrievable IVC filters demonstrated an average filter retrieval rate of merely 37%, with most filter-related complications occurring >30 days after placement.

对于抗凝或其他导致静脉血栓栓塞高风险的因素有禁忌症的患者，腔静脉滤器长期以来一直是一种预防肺栓塞的方法。一般来说，与潜在的可致毁灭性发病率和死亡率的肺栓塞相比，腔静脉滤器是相对安全的医疗器具。然而滤器并非完全没有风险，特别是本该取出的滤器。对37项研究和6834个可回收腔静脉滤器的系统回顾表明，平均滤器回收率仅为37%，大多数滤器相关并发症发生在放置后30天以后。

图5. 按滤器位置分层的回收失败率。HA，钩/顶点；HA+C，钩/顶点和滤器衣领部

The overall effectiveness of IVC filters in reducing morbidity and mortality has continued to be debated. Filters have been demonstrated to reduce the risk of PE but have also been associated with an increased risk of proximal DVT.2,11,12Studies of filter effectiveness are lacking, partially owing to the question of equipoise in conducting such studies. However, current evidence has suggested that the overall morbidity and mortality might not be significantly affected by the use of IVC filters.

滤器在降低发病率和死亡率方面的总体效果仍有争议。滤器已被证明可以降低肺栓塞的风险，但也与髂股静脉血栓增加的风险有关。滤器有效性的研究仍缺乏，部分由于在这样的研究中的均衡问题。然而，目前的证据表明，使用腔静脉滤器可能不会对总的发病率和死亡率产生显著影响。

No significant complications or patient morbidity were associated with any of the 295 filter retrieval attempts in the present study, suggesting that this procedure is safe and should be attempted for any patient with an IVC filter that is no longer needed. Although the risks associated with a long filter dwell time are low, the risks in attempting retrieval appear to be minimal. Thus, it would be prudent to attempt retrieval for all patients once the filter is no longer needed. Although permanent filters are designed to be left within the patient indefinitely, 13 retrievals were attempted for permanent filters within our cohort. The decision to remove these filters was at the discretion of the surgeon and retrieval had generally been attempted because the patient no longer had any indication for the filter to remain in place. Complex filter retrieval techniques included snaring the collar of the filter instead of the hook, using endobronchial forceps to grasp the filter, and ballooning the filter away from the caval wall. These techniques were often used during the more difficult filter retrieval attempts, once simple retrieval by snaring the hook of the filter had failed.

在本研究中295例滤器取出尝试中，没有与操作相关的并发症，表明该操作是安全的，所以对于任何不再需要滤器的患者都应该尝试取出。虽然滤器长时间体内停留的风险很低，但尝试取出的风险似乎很小。因此，当不再需要滤器时，尝试回收所有患者体内滤器是谨慎推荐的。尽管永久性滤过器被设计为无限期地留在患者体内，但在我们的队列中尝试了13次永久性滤器的取出。当没有继续将滤器留置在患者体内的适应症时都应尝试取出，取出这些滤器的决定是由外科医生决定的。复杂的滤器回收技术包括圈套滤器领口部，用支气管内镜活检钳取滤过器，球囊扩张使之与血管壁剥离。这些技术通常在常规回收操作失败后的更困难的滤器回收尝试中使用。

图6. 按年份分层的滤器回收失败率

Tine penetration through the caval wall was nearly ubiquitous within our patient cohort and, thus, could not be used as a predictor of filter retrieval failure. However, the position of the hook/apex and collar relative to the caval wall can be very useful in predicting whether filter retrieval will be successful. Association of the filter hook/apex itself or the filter hook/apex and collar with the caval wall has been significantly associated statistically with an increased rate of retrieval failure and, thus, can be used as a predictive factor when assessing the likelihood of procedural success. Association of any aspect of the filter with nearby structures, including the bowel, liver, and spine, was not associated with an increased rate of retrieval failure.

滤器齿穿透腔静脉壁的情况在我们的患者队列中很常见，因此，不能作为滤器回收失败的预测因素。然而，回收钩/尖端和领口部相对于管壁的位置对于预测滤器回收是否成功非常有用。回收钩/顶端本身（HA）或回收钩/顶点和领口部（HA+C）与腔静脉壁的相关性在统计学上与回收失败率的增加显著相关，因此，在评估手术成功的可能性时，可作为预测因素。滤器的任何部分与邻近结构（包括肠、肝和脊柱）的关联与回收失败率的增加无关。

A total of 10 patients (3.4%) in the present study had had a filter placed despite the lack of a clear indication for placement, such as a contraindication to anticoagulation therapy. Filter placement outside the recommended indications has also been demonstrated in other studies.13Although this trend accounted for only a small portion of our patient cohort, it is still concerning, because filters should not be placed without a clear indication. Care should be taken when assessing patients for the possibility of filter placement, and filters should only be placed when the potential benefits outweigh the possible risks.

本研究共有10名患者（3.4%）在没有明确的放置指征（如抗凝治疗禁忌症）的情况下放置了滤器。在推荐适应症之外的滤器放置也在其他研究中也有报道。尽管这一趋势只占我们患者队列的一小部分，但它仍然令人关注，因为没有明确的适应症就不应放置滤器。在评估患者放置过滤器的可能性时应小心，只有当潜在的益处大于可能的风险时才应放置滤器。

In recent years, placement of IVC filters has appeared to be decreasing modestly, although the rate of IVC filter retrieval attempts has been increasing yet has remained low. Imperative to the goal of decreasing IVC filter complications is the need for improved rates of retrieval, because most complications have occurred in filters that have remained indwelling longer than would be ideal. The retrieval rates might be improved through implementation of a multidisciplinary protocol, which has previously been demonstrated at another institution.

近年来，尽管腔静脉滤器的取出尝试率一直在增加，但滤器的放置似乎在适度减少。降低滤器并发症的目标必须提高取回率，因为大多数并发症发生在滤器留置的时间比理想的要长。通过实施一个多学科会诊方案，回收率可能会有所提高，这一点以前已经在另一个机构得到证实。

Our study was limited by the inherent nature of retrospective reviews. Some data were limited for collection, especially because a subset of patients had had their filter placed at an outside institution. The approach to the evaluation and technique was developed throughout the study period. Additionally, we performed a single-institution study, potentially limiting the ability to generalize the results to the population at large. We also had a low number of permanent filters compared with those designed to be temporary to study the differences between these groups. Additionally, our institution likely treats a skewed number of patients with malpositioned filters and long filter dwell times. We also did not include filter tilt in our analysis, although previous studies have suggested that filter tilt might have a significant effect on filter retrieval success. Future directions could include implementation of a filter retrieval protocol to ensure that the device will be removed as soon as it is no longer medically indicated. Such a protocol would be a relatively simple method to decrease the complications associated with IVC filters and has been supported by the results of the present study, which indicated that retrieval should be attempted for all filters, given the demonstrated procedural safety.

我们的研究受到回顾性研究固有性质的限制。一些数据的收集受到限制，特别是因为有一部分患者在外院放置的腔静脉滤器。评估方法和技术是在整个研究期间发展起来的。此外，我们做的是一项单中心研究，这可能会限制将结果推广到广大人群的能力。另外，和临时滤器数量相比我们永久性滤器数量偏少，不足以研究亚组之间的差异。此外，我们的机构回收了可致研究数据偏倚数量的移位滤器和超回收窗滤器。我们也没有在我们的分析中包括滤器倾斜，尽管先前的研究表明滤器倾斜可能对滤器回收成功有重大影响。未来的方向可能包括实现滤器回收流程，以确保临床不再需要时被移除。这样一个方案将是一个相对简单的方法，以减少与滤器相关的并发症，并已得到本研究结果的支持，即考虑到已被证明的操作安全性，应尝试对所有滤器进行回收。

CONCLUSIONS结论

The results of the present study have suggested that retrieval should be attempted for all IVC filters, irrespective of the chronicity and complexity, given the procedural safety. In our study, tine penetration was nearly ubiquitous. However, IVC filters with the HA or HAtC imbedded into or penetrating through the caval wall are predictors of retrieval failure. Obtaining a preoperative CT scan can be helpful in assessing the difficulty or complexity of retrieval. Improved rates of filter retrieval should be attempted to decrease the risks associated with prolonged filter dwell times. We encourage the use of IVC filter retrieval protocols as a relatively simple method to limit the filter dwell times.

本研究的结果表明，无论程序的长期性和复杂性如何，都应尝试对所有滤器进行回收，因为操作安全。在我们的研究中，滤器齿穿透血管壁几乎无处不在。然而，回收钩（HA）或回收钩和滤器领口部（HA+C）嵌入或穿透腔壁是回收失败的预测因素。术前CT扫描有助于评估回收的难度或复杂性。应尝试提高滤器回收率，以减少与延长滤器体内停留时间相关的风险。我们鼓励使用滤器回收多学科会诊流程作为一种相对简单的方法来减少滤器体内停留时间。

AUTHOR CONTRIBUTIONS 作者贡献（略）

Conception and design: KM, JB, SL, EP, CS

Analysis and interpretation: KM, JB, CS

Data collection: KM

Writing the article: KM Critical revision of the article: KM, JB, SL, EP, CS

Final approval of the article: KM, JB, SL, EP, CS

Statistical analysis: JB

Obtained funding: Not applicable

Overall responsibility: KM

REFERENCES参考文献（略）