Prophylactic protocol for cerebral infarction after left upper lobectomy and prospective analysis of pulmonary vein stump thrombosis

Introduction

Postoperative cerebral infarction (CI) is a rare complication of lung resection that can cause paralysis or other neurological deficits and impair the patient’s quality of life. Reported incidence rates range from 0.4% to 1.1%. The incidence of CI is higher after left upper lobectomy (LUL) than after resection of other lobes (1-3). Up to 30% of CIs occur within 7 days of surgery; their frequency decreases dramatically after 30 days and plateaus after 4 months (4). Prevention of CI is necessary in the first week after surgery in patients at high risk.

One of the causes of CI after lung resection is thought to be thrombus in the pulmonary vein (PV) stump (PVS) or cardiac atrium. This is based on the hypothesis that the PVS causes blood flow stagnation, which triggers thrombus formation (5-7). PVS thrombosis (PVST) is more frequently detected after LUL than after right-sided anatomical resection (30.8% vs. 4.9%), and 95% of PVST events occur within 1 week of surgery (8). Shortening the PV stump via proximal ligation can prevent CI after LUL (9-12). Although atrial fibrillation (AF) is also known to cause stagnant blood flow, leading to the generation of thrombosis (13), no measures of prevention in the post-lobectomy clinical setting have yet been established. In our institution, a protocol for preventing CI was introduced on April 20, 2021. This study aimed to identify risk factors for PVST and investigate the safety and efficacy of our CI prevention protocol. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2268/rc).

Methods

Study design and patients

This single-center observational study compared patients who underwent LUL with the CI prevention protocol we initiated on April 20, 2021, with those treated during the preceding 5 years (between January 1, 2016 and April 19, 2021). The protocol was as follows: (I) ligate the left superior PV as proximal as possible during the operation; (II) postoperative nursing neurological examinations every 4 hours in the first 72 hours after surgery; (III) administration of subcutaneous heparin (5,000 units twice daily) for 7 days after surgery; (IV) electrocardiographic (ECG) monitoring for 7 days after surgery; and (V) contrast-enhanced computed tomography (CECT) on postoperative day 7 to examine for PVST—if present, initiate anticoagulation and perform follow-up CECT 4 months later. Patients with diabetes mellitus and hemoglobin A1c ≥8% were hospitalized and treated by a diabetologist before surgery. Smokers were required to quit smoking for at least 1 month before surgery. Patients requiring bronchoplasty or angioplasty were excluded from the study.

Among the 148 patients who underwent LUL in the 5 years before protocol implementation, four (2.7%) developed CI, and two of these patients died. On the basis of this incidence rate, we established an incidence threshold of 5% and an expected incidence rate of 0% to conclude that the newly introduced protocol was effective. With a one-sided alpha of 5% and 90% power, the required sample size was calculated to be 59 patients. Therefore, 73 patients were enrolled after protocol implementation (from April 20, 2021 to December 2024) to account for possible loss to follow-up. All data were collected prospectively and recorded.

Details of surgical procedures and postoperative monitoring in the protocol

Ligation of the PV was performed as proximal as possible outside the pericardium using a 1-0 silk suture before PV division or an endoscopic loop device (Endoloop, Ethicon, Raritan NJ, USA; or SurgiTie, Covidien, Dublin, Ireland) after PV division. If PV ligation was difficult or impossible to perform safely, it was divided without ligation at the surgeon’s discretion. A neurological examination was performed by nursing staff every 4 hours for 72 hours after surgery to closely monitor for the development of CI symptoms. Subcutaneous heparin was introduced on the evening of the day of surgery (5,000 units twice daily), which was based on the venous thromboembolism prophylaxis dosage recommended by the American Heart Association (14). Heparin was discontinued in those originally taking oral anticoagulants after the oral medication was resumed. Delaying or discontinuing heparin administration was permitted if the operating surgeon deemed the patient to have a high risk of postoperative bleeding based on intraoperative findings, such as excessive bleeding. If AF was detected on ECG, antiarrhythmic agents were introduced at a cardiologist’s discretion. CECT was not performed on postoperative day 7 or during follow-up in patients with renal dysfunction or allergy to contrast media. In the patients who were diagnosed with PVST on CECT, a direct oral anticoagulant was introduced according to the second update to the ninth edition of the CHEST guidelines for treatment of deep vein thrombosis (15), and follow-up CECT was performed 4 months after surgery.

Outcomes

The incidence of CI within 4 months of surgery, incidence of postoperative AF (including both persistent AF and temporary AF) in the first 7 days after surgery, and postoperative complications were compared between patients treated during the 5 years preceding the protocol and those treated under the protocol. The incidence of PVST and its risk factors were analyzed in patients who underwent CECT on postoperative day 7. Postoperative complications were assessed using the Clavien-Dindo classification (16). The length of the PV pocket was measured from the left superior PV stump to the left atrium on coronal CECT imaging on day 7 with 0.5 mm slices in the patients who underwent CECT in the protocol. Measurements were performed by a single reviewer. The view that showed the longest PV pocket was used (Figure 1).

Figure 1 Images of pulmonary vein stump after left upper lobectomy are shown in two modalities. (A) Coronal image of contrast-enhanced computed tomography. The length of pulmonary vein pocket was measured between the red dotted lines. (B) Three-dimensional image on contrast-enhanced computed tomography after left upper lobectomy. The length of pulmonary vein pocket is indicated by the white dotted lines.

Statistical analysis

Continuous variables are presented as means with range and were compared using the unpaired t-test. Categorical variables are presented as frequencies with percentage and were compared using the Chi-squared test. For the outcomes assessed in this study, adjusted odds ratios (ORs) were estimated using Firth’s penalized logistic regression (17), accounting for potential temporal confounding due to changes in surgical approach and operative time over the study period, to address sparse data and low event rates. When complete separation occurred (zero events in one group), Haldane-Anscombe continuity correction (+0.5 to all cells) was additionally applied to 2×2 contingency tables (18). 95% confidence intervals for event rates were calculated using Wilson score intervals (Rule of 3 for zero events) (19). In addition, risk factors associated with postoperative PVST were analyzed using Firth’s penalized logistic regression to account for the low event rates. The model included patient background (age, comorbidities, treatment with antiplatelet or anticoagulant drugs, preoperative chemotherapy) and postoperative course (postoperative AF, length of PV pocket) as associated factors (2,4). Statistical analyses were performed using JMP Student Edition 18 (JMP Statistical Discovery LLC, Cary, NC, USA), a graphical user interface for R software (www.r-project.org).

Ethical statement

The study was approved by the Kyoto University Certified Review Board (No. R2750, on December 2020). All patients provided written informed consent. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Results

Patient characteristics

Patient characteristics and perioperative findings in the protocol and before protocol groups are shown in Table 1. The prevalence of robot-assisted thoracic surgery (RATS) was higher [19 cases (26.0%) vs. 16 cases (10.8%), P<0.01] and mean operation time was longer (238 vs. 211 minutes, P<0.01) in the protocol group. Among the 73 patients in the protocol group, ligation of the PV was performed as proximally as possible outside the pericardium in 50 patients (68.4%). There were various reasons for not performing PV ligation. Anatomical reasons included the PV being thick and short, as well as branching proximally. Other reasons included difficulty in performing proximal ligation as a result of dividing the PV with staples. During RATS, the frequency of left upper PV ligation was relatively low (3/19, 15.7%), which was because the left upper PV was divided using robot staplers. In cases in which the PV had been divided with a stapler beforehand, further proximal ligation of the vein was technically challenging.

All patients with a pre-existing diagnosis of AF received anticoagulation therapy. Postoperative management of these patients included heparin administration followed by prompt resumption of their preoperative anticoagulant regimen after the chest drain was removed. Of the 9 patients (12.3%) who did not receive subcutaneous heparin in the protocol group, treatment was withheld in four owing to concerns about rebleeding based on intraoperative findings. The remaining five patients were managed using alternative anticoagulant therapy: two promptly resumed their original regimen, and three received intravenous heparin.

Postoperative outcomes

The incidence rates of CI and other complications are shown in Table 2. Postoperative CI within 4 months of surgery did not occur in the protocol group (0/73 cases: 0.0%, 95% confidence interval: 0.0–4.1%); in contrast, the incidence of postoperative CI in the before protocol group was 2.7% (4/148 cases, 95% confidence interval: 0.1–5.3%; adjusted OR 0.15, 95% confidence interval: 0.00–1.57). Postoperative AF occurred in nine patients in the protocol group and 16 patients in the before protocol group (12.3% and 10.8%, respectively; adjusted OR 0.90, 95% confidence interval: 0.36–2.12). In the protocol group, new-onset AF occurred on postoperative day 1 in two patients, day 2 in three, and day 5 in one. All were temporary AF. Among these, three were asymptomatic and detected by postoperative ECG monitoring. Among the four patients with a history of AF, postoperative ECG monitoring detected AF in three cases: one had chronic persistent AF, and two had temporary recurrent episodes on day 3 and day 5, respectively.

Regarding postoperative complications, the incidence of grade 3 or higher complications did not significantly differ between the protocol and before protocol groups (16.4% vs. 14.9%; adjusted OR 0.89, 95% confidence interval: 0.38–1.98). No postoperative hemorrhage or complications related to the CI prevention protocol were observed. No patient required re-operation or postoperative replacement of a chest tube.

Of the 73 patients in the protocol group, 64 (87.7%) underwent postoperative CECT, and PVST was detected in five (5/64=7.8%). Among the nine patients in whom CECT was not performed, the reason was renal dysfunction in two, contrast allergy in four, asthma in one, and patient preference in one. Among the five patients with PVST, CECT 4 months after surgery showed resolution of the thrombus in three and a reduction in thrombus size in one; CECT was not performed in the remaining patient owing to poor renal function caused by adjuvant chemotherapy. After thrombus resolution, anticoagulation therapy was discontinued in the patients without postoperative AF, followed by no late onset CI until this analysis (Figure 2).

Figure 2 In all patients diagnosed with pulmonary vein stump thrombosis, contrast-enhanced computed tomography revealed an area of contrast defect at the site of the pulmonary vein stump. Anticoagulation therapy was initiated in all, and repeat imaging 4 months later showed improvement in the contrast defect.

PVST risk factors

Patient characteristics overall and in patients with and without postoperative PVST are shown in Table 3. Because two patients with postoperative AF did not undergo CECT owing to poor renal function or contrast allergy, the subsequent analysis was conducted in the remaining seven patients with AF (three with chronic AF, and four with new-onset AF). Age, length of PV pocket, and prevalence of postoperative AF significantly differed between the PVST and no PVST groups. Multivariable logistic regression analysis showed that length of PV pocket (OR, 1.29; 95% confidence interval: 1.05–1.84; P=0.01) and postoperative AF (OR, 15.18, 95% confidence interval: 1.54–334.99; P=0.02) were independent risk factors for PVST (Table 4). Furthermore, the distribution of PV pocket according to PVST status was examined (Figure 3), and four of the five patients with PVST had a PV pocket >2.5 cm. Therefore, we examined a model in which PV pocket >2.5 cm was included as an explanatory variable (Table 5). This model found that PV pocket >2.5 cm was an independent risk factor for PVST (OR, 22.6; 95% confidence interval: 2.17–1120.45; P=0.01).

Figure 3 Distribution of length of pulmonary vein pocket in patients with and without postoperative pulmonary vein stump thrombosis. PV, pulmonary vein; PVST, pulmonary vein stump thrombosis.

Discussion

This study aimed to evaluate the efficacy and safety of our prophylactic protocol for postoperative CI after LUL. The protocol was based on findings in previous studies and focused on PV ligation (9-12), postoperative AF (13), and the hypothesis that PVST may cause CI after lung resection (8). As a result, no instances of CI, postoperative bleeding, or severe complications occurred. Although postoperative CI is a rare complication, we calculated that 59 cases were necessary to demonstrate the effectiveness of the protocol based on past data from our institution. In this study, no CI events were observed among 73 patients in the protocol cohort (0.0%, 95% confidence interval: 0.0–4.1%). This narrow interval suggests a low absolute risk under the protocol, but also indicates that non-zero risks up to approximately 5% cannot be excluded and that the findings should be interpreted with caution. In addition, vascular damage during PV ligation and bleeding after postoperative heparin administration are expected complications, and in patients in whom the risk of complications is high, it was acceptable not to perform one of the preventive measures by combining multiple preventive measures. Operation time and surgical approach significantly differed between the protocol and before protocol groups, and it cannot be ruled out that these differences may have influenced the outcomes. RATS reportedly has longer operation times and lower rates of cardiovascular complications (20). In this study, the RATS approach was more frequent in the protocol group, which may have influenced the results; therefore, we cannot definitively state that our protocol was effective. However, the observation that no cases of CI occurred after protocol initiation is considered noteworthy. Furthermore, no differences were observed between the protocol and before protocol groups regarding traditional thrombotic risk factors such as AF, hypertension, hyperlipidemia, CI, or diabetes.

We also prospectively examined the incidence and risk of PVST, which is thought to be a cause of CI after lung resection. There is no clear evidence regarding the timing of PVST formation, and consequently, no definitive basis for the timing of CECT imaging. However, previous reports indicate that up to 30% of CI cases occur within the first 7 days after surgery (4). In this protocol, postoperative observation is conducted for 1 week, with discharge decisions made following CECT evaluation. Therefore, CECT imaging is scheduled for postoperative day 7. Hattori et al. reported that 30.8% of patients who underwent LUL developed PVST (8). In our study, the incidence of PVST was 7.8%, which is considerably lower. This marked reduction may reflect the benefit of the preventive regimen. We selected anticoagulant therapy as the prophylactic medication, based on the hypothesis that the PVS causes blood flow stagnation, which triggers thrombus formation (5-7). Subcutaneous heparin was specifically chosen because it is easy to implement for both patients and medical staff, and the dosage was determined based on standard well-established venous thromboembolism prophylaxis recommendations. Although intravenous heparin and oral anticoagulants are considered more aggressive preventive measures, given that their preventive efficacy against PVST has not yet been established, we ascertained that the bleeding risk associated with both measures was too high. In all five patients with PVST, aggressive anticoagulant therapy was initiated based on the treatment protocol for venous thromboembolism, resulting in improvement on follow-up imaging at 4 months. In a study of CI after pulmonary resection in Japan, the frequency of postoperative CI decreased dramatically after 30 days and plateaued after 4 months (4). Anticoagulant therapy for pulmonary embolism is recommended for at least 3 months (21). Therefore, in patients in whom improvement in the thrombus was confirmed, anticoagulant therapy was discontinued (provided they did not experience postoperative AF), and no late-onset CI was observed. Prediction of PVST and the appropriate timing of evaluation and treatment are considered important.

Several studies have reported that a long PV pocket causes blood flow to stagnate, which may lead to thrombus formation (5-7). The effectiveness of shortening the PV pocket by proximal PV ligation has also been demonstrated (9-12). In our multivariable analysis, PV pocket length was associated with PVST, which agrees with these previous reports. However, in some patients, the PV pocket was 2.5 cm or longer despite PV ligation. This may have occurred because the length of the PV pocket depends on the length of the PV within the pericardium. One solution is to ligate the PV within the pericardium. However, PV ligation was difficult or considered high risk in 31.5% of the patients in our study, raising concerns about the safety of pericardial ligation. In robotic surgery specifically, the rate of PV ligation was relatively low. In our institution, PVs are transected proximally using a stapler during RATS, which may have led many surgeons to consider PV ligation at a more proximal site as high risk. Alternatively, the distance between the assistant port and the left superior PV may have been too far, making ligation of the PV from the assistant port difficult. It is possible that extra-pericardium proximal ligation of the PV alone is not sufficient for preventing CI in patients with a long intrapericardial PV segment; such patients may require additional prophylaxis measures.

Postoperative AF was also a risk factor for PVST. It was detected in 12.3% of patients in the protocol group, and the incidence of PVST in this subgroup of patients was 42.8%. Among the nine patients with postoperative AF, it was new-onset in six, and three were asymptomatic. Five of the six new-onset cases occurred within the first 2 days after surgery. The presence of postoperative AF was associated with the development of PVST and may have increased the risk of postoperative CI. Kimura et al. (2) reported no association between CI and postoperative arrhythmia; however, asymptomatic AF might have been overlooked because it was a database study. Fifty percent of new-onset cases in our study were asymptomatic; therefore, ECG monitoring is recommended in the early postoperative period to avoid missing these cases. Considering the high risk of PVST, CECT is strongly recommended for patients who develop postoperative AF. In cases where CECT cannot be performed, other imaging studies such as cine magnetic resonance imaging should be considered. Furthermore, postoperative AF frequently occurred in the early postoperative period, so earlier implementation of CECT, for example within the first 3 days after surgery, might enhance its effectiveness to prevent postoperative CI caused by PVST.

This study has several limitations. First, it was conducted in a single center and the sample size was small. Especially in multivariable analysis, there were only five cases of PVST. Multicenter large-scale studies are needed to more accurately demonstrate efficacy. Second, the protocol and before protocol groups differed in a few background factors, including temporal changes in surgical practices. Although we adjusted for surgical approach and operative time—key factors that differed between periods—residual confounding from unmeasured temporal factors (e.g., team experience, anesthetic protocols, changes in perioperative management) cannot be excluded, and the protocol effect cannot be completely separated from background secular trends in this historical control design. Third, nine patients were excluded from PVST analysis because they did not undergo CECT; two of these patients had postoperative AF, so the impact of postoperative AF may have been underestimated. The fact that CECT was not performed in all protocol group patients may have introduced selection bias. Furthermore, although multivariable analysis was performed to evaluate known risk factors, there may have been other potential risk factors that were not included in the model. Additionally, low event rates across outcomes resulted in wide confidence intervals despite using Firth’s penalized logistic regression and Haldane-Anscombe correction for sparse data, indicating residual statistical instability. Future studies should investigate additional risk factors and employ prospective randomized designs to establish causality.

Conclusions

In conclusion, after our protocol for preventing CI after LUL was implemented, no CI within 4 months of surgery or serious complications was observed. PVST occurred in 7.8% of patients, and PV pocket length and postoperative AF were risk factors associated with PVST. Early postoperative CECT is useful for the early detection and management of PVST, and it is especially recommended in patients with postoperative AF.

Acknowledgments

We thank Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2268/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2268/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2268/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2268/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was approved by the Kyoto University Certified Review Board (No. R2750, on December 2020). All patients provided written informed consent. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Yamamoto T, Suzuki H, Nagato K, et al. Is left upper lobectomy for lung cancer a risk factor for cerebral infarction? Surg Today 2016;46:780-4. [Crossref] [PubMed]
2. Kimura D, Yamamoto H, Endo S, et al. Postoperative cerebral infarction and arrhythmia after pulmonary lobectomy in Japan: a retrospective analysis of 77,060 cases in a national clinical database. Surg Today 2023;53:1388-95. [Crossref] [PubMed]
3. Hattori A, Takamochi K, Kitamura Y, et al. Risk factor analysis of cerebral infarction and clinicopathological characteristics of left upper pulmonary vein stump thrombus after lobectomy. Gen Thorac Cardiovasc Surg 2019;67:247-53. [Crossref] [PubMed]
4. Matsumoto K, Sato S, Okumura M, et al. Left upper lobectomy is a risk factor for cerebral infarction after pulmonary resection: a multicentre, retrospective, case-control study in Japan. Surg Today 2020;50:1383-92. [Crossref] [PubMed]
5. Ohtaka K, Takahashi Y, Uemura S, et al. Blood stasis may cause thrombosis in the left superior pulmonary vein stump after left upper lobectomy. J Cardiothorac Surg 2014;9:159. [Crossref] [PubMed]
6. Ichimura H, Ozawa Y, Nishina H, et al. Thrombus formation in the pulmonary vein stump after left upper lobectomy: a report of four cases. Ann Thorac Cardiovasc Surg 2014;20:613-6. [Crossref] [PubMed]
7. Nakano T, Kaneda H, Murakawa T. Stagnating blood flow related to thrombus formation in pulmonary vein stump after left upper lobectomy. Gen Thorac Cardiovasc Surg 2023;71:648-56. [Crossref] [PubMed]
8. Hattori A, Takamochi K, Shiono S, et al. Multicentre prospective observational study for pulmonary vein stump thrombus after anatomical lung resections. Eur J Cardiothorac Surg 2021;61:92-9. [Crossref] [PubMed]
9. Watanabe SN, Imai K, Mitsui M, et al. Proximal ligation technique prevents thrombus formation in the pulmonary vein stump after lobectomy. Surg Today 2024;54:839-46. [Crossref] [PubMed]
10. Mizukami Y, Tada M, Adachi H. Cerebral infarction after left upper lung lobectomy with central vascular ligation. J Thorac Dis 2020;12:2210-9. [Crossref] [PubMed]
11. Nakano T, Kaneda H, Kawaura T, et al. Ligating the pulmonary vein at the pericardial reflection is useful for preventing thrombus formation in the pulmonary vein stump after left upper lobectomy. Gen Thorac Cardiovasc Surg 2019;67:450-6. [Crossref] [PubMed]
12. Nishimura T, Okada S, Kaneyama K, et al. Proximal ligation of the pulmonary vein stump to prevent postoperative cerebral infarction after a lobectomy. Eur J Cardiothorac Surg 2025;67:ezaf041. [Crossref] [PubMed]
13. Yamashita T. Virchow triad and beyond in atrial fibrillation. Heart Rhythm 2016;13:2377-8. [Crossref] [PubMed]
14. Hirsh J, Anand SS, Halperin JL, et al. Guide to anticoagulant therapy: Heparin : a statement for healthcare professionals from the American Heart Association. Circulation 2001;103:2994-3018. [Crossref] [PubMed]
15. Stevens SM, Woller SC, Kreuziger LB, et al. Antithrombotic Therapy for VTE Disease: Second Update of the CHEST Guideline and Expert Panel Report. Chest 2021;160:e545-608. [Crossref] [PubMed]
16. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205-13. [Crossref] [PubMed]
17. Firth D. Bias reduction of maximum likelihood estimates. Biometrika 1993;80:27-38.
18. Haldane JBS. The estimation of the probabilities of the causes of a disease. Am Stat Assoc J 1956;51:100-10.
19. Wilson EB. Probable inference, the law of succession, and statistical inference. J Am Stat Assoc 1927;22:209-12.
20. Catelli C, Corzani R, Zanfrini E, et al. RoboticAssisted (RATS) versus Video-Assisted (VATS) lobectomy: A monocentric prospective randomized trial. Eur J Surg Oncol 2023;49:107256. [Crossref] [PubMed]
21. Kahn SR, de Wit K. Pulmonary Embolism. N Engl J Med 2022;387:45-57. [Crossref] [PubMed]