Arterial stump thrombosis (AST) in pulmonary resection surgery has been described, but this condition, particularly the prognostic implications and the correct therapeutic approach, has not been studied in depth. Since the first 2 cases of AST in pneumonectomies were described in 1966,1 only 2 studies,2,3 small series and isolated case reports have been published in the literature, all retrospective. Although in the past little importance has been given to this finding, in recent years, the generalized use of imaging techniques in follow-up, and the improved resolution of these procedures have led to a growing number of reports of cases with associated complications.4–11

The main objectives of our study were to determine the prevalence of AST in oncological lung resection surgery, to identify risk factors present before the intervention and those caused by the procedure itself, and to describe subsequent progress in terms of radiological follow-up and treatment. The secondary objective was to determine clinical, oncological and prognostic differences between patients with AST, patients who did not have any type of thrombosis, and those who had pulmonary thromboembolism (PTE).

During the study, 507 thoracotomies were performed, of which 19 cases (4%) were excluded because they were segmentectomies; 9 patients (1.8%) were excluded because they died in the first 2 weeks after the intervention; and 6 (1.2%) were excluded because no CTA with intravenous contrast medium was performed during follow-up. Thus, a total of 473 surgeries were included, with final diagnoses of lung cancer in 471 cases (99.5%), cutaneous T cell lymphoma with pulmonary involvement in 1 patient and fungal infection in another. Resection type was lobectomy in 393 (83%), pneumonectomy in 61 (13%), and simultaneous bilobar resections in 19 (4%). The most common intervention type was right upper lobectomy (147 patients, 31%) and the most common histological strain was squamous cell carcinoma (213 patients, 45%). A total of 388 patients were men (82%), and mean age was 65.3±9.5 years. The patient selection flow chart is shown in Fig. 1 and other baseline characteristics in Table 1.

Fig. 1.

Patient selection flow chart. AST: arterial stump thrombosis; CTA: computed tomography angiogram; PTE: pulmonary thromboembolism.

(0.15MB).

PET was diagnosed in 10 cases (2.1%) and AST in 9 (1.9%) (Fig. 2), resulting in an incidence of 1.9 cases of suture thrombosis per 100 patients operated. Table 2 shows surgical, radiological and oncological data by patient groups. Median follow-up after surgery was 49.4 months (interquartile range 34.8–70.1 months), and 147 deaths were documented. Table 3 describes causes for death by patient groups.

Fig. 2.

Multidetector computed tomography axial slice after intravenous contrast in arterial phase in patients with history of right pneumonectomy. A filling defect consistent with arterial stump thrombosis (black arrow) and associated sutures (white arrow) are seen.

(0.08MB).

After surgery, median time until AST was visualized was 11.3 months (interquartile range 2.7–42.2 months) with a range of 67.5 (1.4–69.0 months); of the 9 cases, 5 (56%) occurred within the 12 months following surgery. Seven (78%) of the AST patients began antiplatelet treatment. Three patients (33%) died, 2 of whom were receiving antiplatelets. Of these, 1 had a lymphoproliferative disorder, and 1 had an acute subdural hematoma while also receiving chemotherapy for tumor relapse. Median duration of antiplatelet treatment was 298 days (interquartile range 45–532 days). Filling defects were resolved, according to radiological findings, in 6 cases (67%), 2 of which were not receiving antiplatelets (22%). Media time to resolution was 12.9 months (interquartile range 3.8–20.3 months). Mean number of CTAs performed during follow-up of AST patients was 11.7±6.3 CTAs, statistically higher than the mean in patients without AST or PTE, who had 6.9±4.9 CTAs (P=0.02).

Follow-up after appearance of AST was 44 months (interquartile range 11.8–56.0 months). The Kaplan–Meier survival analysis revealed a greater accumulated survival in patients who did not present PTE or AST during follow-up, with a mean of 80.9 months (95% CI: 76.6–85.2 months), while patients with AST had a median survival of 67.4 months (95% CI: 47.4–87.4), and patients with PTE had the shortest survival at 56.9 months (95% CI: 42.1–71.8). However, none of the differences was statistically significant (log-rank 1.148, P=.56) (Fig. 3).

Fig. 3.

Kaplan–Meier subgroup survival analysis: pulmonary thromboembolism (thin continuous line), arterial stump thrombosis (thick continuous line), and patients without thromboembolism or arterial stump thrombosis (dotted line).

(0.07MB).

Finally, surgical, clinical and oncological characteristics of patients who developed AST, divided into subgroups of “within 12 months” following surgery and “after 12 months” following surgery, are presented in Table 4.

The most important result of our study was the determination of the epidemiological and clinical features of AST in lung resection surgery. AST was seen to be rare, with a tendency to occur in surgeries performed in the right side. It was not associated with pre-surgery risk factors, and was unrelated to histological strain, oncological staging or tumor relapse. Appearance over time was variable and it did not impact on patient survival.

With reference to prevalence, Kim et al.2 reported a total of 18 cases of AST in 147 pneumonectomies (12.2%), practically the same as that found by Kwek and Wittram3 who observed 11 clots in 89 patients (12.4%) also subjected to pneumonectomies. These authors focused their studies on the correlation between stump length and changes in blood flow dynamics, and included only pneumonectomies. Nevertheless, the prevalence of AST in this subgroup in our series is still low, with 2 cases among 61 patients (3.3%). The main reason for these different findings may be technical, since continuous suture was performed until the first semester of 2010, after which mechanical sutures with inert metal staples were used. A second reason may be compliance with a strict anticoagulation protocol associated with the surgery type and risk of VTED, proposed by the Spanish Society of Pulmonology and Thoracic Surgery.13 Our results, then, are closer to those of studies which have analyzed the prevalence of VTED in thoracic surgeries in general. In one series of almost 700 patients, a prevalence of 1.3% was determined, with no mention being made of AST,17 and in another Spanish series, a 0.18% prevalence was reported among 6,004 patients undergoing any type of chest surgery, with 7 cases of PTE and 4 of AST. It is worth noting that all cases of PTE had an underlying cancer diagnosis, and 6 of them had undergone major surgery, including pneumonectomies, lobectomies or atypical segmentectomies.18 With regard to the surgical technique, Işik et al.19 found that clot formation and embolization were more likely on transfixation sutures of left pneumonectomies, compared to continuous sutures.

The type of vessel sutured has been previously studied. While reports of embolic complications and infarctions in the case of AST are anecdotal, clots on venous sutures have been significantly associated with infarction, particularly in the brain. Recently, Yamamoto et al.20 presented a series of 6 patients who underwent left upper lobectomies complicated by cerebral infarction, of which 3 cases of clot on the corresponding vein were documented. Intraoperative ultrasonography has also shown that greater hemostasis occurs in the venous stump when the surgical procedure is left upper lobectomy.21 No infarction of vital organs was found in our series.

The association between the side of surgery and the appearance of AST has not been clearly determined. In contrast to the 2 studies mentioned above, our AST cases occurred more frequently in surgeries in the right side, as has also been reported in previous studies of pneumonectomies and AST [19 of the 29 cases detected (66%)].2,3 Moreover, in the small series and isolated clinical cases, right-side AST was observed in 6 of 7 cases of pneumonectomy presented (86%)4,5,7–9 and 1 case after right lower lobectomy.6 Following this line of argument, pneumonectomy as a major surgical procedure has been associated with greater activation of the coagulation cascade, and finally, with an increased chance of AST, compared to lobectomies.22

Our study revealed some interesting findings. AST occurring in the first 12 months after surgery was more likely to resolve, even in the 40% of cases which did not receive antiplatelet treatment. In contrast, those detected more than 1 year after surgery had a poor rate of resolution, and long-term antiplatelet treatments were prescribed in all cases. The number of cases is small, so no solid evidence is available on which to base the decision to administer antiplatelets and treatment duration. Adverse events associated with antiplatelet treatment included 1 case of acute subdural hematoma that developed while the patient was receiving active chemotherapy for tumor relapse, so it was impossible to rule out or confirm any association with anticoagulation. Logically, we also found a significantly greater number of CTAs conducted in patients with AST compared to patients without PTE or AST. However, no evidence could be found on the frequency or duration of radiological follow-up, despite the importance of collecting data on a radiological test that is far from risk-free and costly in terms of healthcare budgets.23

Our study has some limitations. Firstly, the loss of follow-up data commonly found in such retrospective studies limits the reliability of our results, although it should be pointed out that to date, none of the studies published on AST has been based on a prospective protocol. Secondly, the low number of cases detected reduces the power of the study and rules out a multivariate analysis for elucidating possible controllable risk factors. Previous series have similar limitations, and no multivariate analyses with more than 2 variables are available due to an insufficient rate of events.2,3 Both studies conclude that stump length is an associated factor. In our series, this measurement could not be determined in all cases as the CTA images obtained before 2008 were not routinely digitalized, and the technical characteristics of the tests were too varied. Thirdly, as mentioned above, this was not a prospective protocol, so patient follow-up was irregular in terms of CTA frequency, particularly during the first 4 years of the study. Moreover, CTAs were evaluated by different doctors, so failure to detect an AST cannot be ruled out.

Our study also has its strengths. This is a large series that included not only pneumonectomies but also other resection procedures, and, being the study with the longest follow-up published to date, it is also the first to present a survival analysis. The clinical relevance of this disease still remains unclear, so this series contributes to our knowledge of the frequency and course of this post-operative event.

To conclude, AST may be no more than an incidental finding, and it is true that its course is benign, but it is equally true that treatment and follow-up remain to be clarified. A larger case series would be desirable to help determine the identification, follow-up and therapeutic approach to this entity.

Daniel López-Padilla participated in data collection and is responsible for the integrity and accuracy of the analysis. He also contributed to the study design, interpretation, analysis and drafting of the final document. Esteban Peghini Gavilanes and Teresa Yolanda Revilla Ostolaza reviewed all the imaging studies to detect arterial stump thrombosis and contributed to the drafting of the final document. María Dolores Trujillo and Iván Martínez Serna retrieved all surgical interventions performed during the study period and contributed to the drafting of the final document. Nuria Arenas Valls, Walther Girón Matute, Roberto Larrosa Barrero, Adriana Manrique Multiozábal, Marta Pérez Gallán and Annette Zevallos collected data and contributed to the interpretation and drafting of the final document. Javier Sayas Catalán contributed substantially to the study design, interpretation, analysis and drafting of the final document.

The authors state that they have no conflict of interests.