Video-assisted mediastinoscopy is safe in patients taking antiplatelet or anticoagulant therapy.

BACKGROUND: The aim of this study was to report our experience with video-assisted mediastinoscopy (VAM) in patients taking antiplatelet (AP) or anticoagulant therapies focusing on perioperative complications (especially haemorrhagic). PATIENTS AND METHODS: We have done a retrospective study from a prospectively maintained database with diagnostic VAM (01/2008-06/2012). We included 54 patients with AP (41 patients - Group A) and anticoagulant (13 patients - Group B) therapies. The control group was formed by 263 patients (Group C). Data regarding the clinical records of the patients, operative time, per- and post-operative complications, total numbers of biopsies and the results of the pathologic examination were collected. We compared the groups A+B versus C, and then A versus C. Statistical differences were calculated by Chi-square test.
RESULTS: In Group A, we had two minor complications: cardiac arrhythmia and peroperative minor haemorrhage. The mean operative time was 29 min and the mean post-operative stay was 1.08 days. In Group B, we had one minor complication: Peroperative minor haemorrhage. The mean operative time was 35 min and the mean post-operative stay was 1.07 days. In Group C, the mean operative time was 28 min. One death occurred (mortality rate of 0.38%) because of cardiac arrest at the induction of anaesthesia. One major complication occurred (severe respiratory insufficiency needing re-intubation) and eight minor complications. Morbidity rate was 2.28%. Mean post-operative stay was 1.14 days. No statistical difference was noted between groups.
CONCLUSION: VAM can be safely performed in patients receiving AP or anticoagulant treatments. There is no increase in peroperative bleeding or post-operative compressive cervico-mediastinal haematoma.

INTRODUCTION

Video-assisted mediastinoscopy (VAM) is used for the diagnosis of superior mediastinal disease and/or staging for lung cancer. Its role has substantially diminished over the past decades because of advances in non-invasive imaging modalities: first, in computed tomography (CT), then in positron emission tomography and more recently, endobronchial ultrasound technique with transbronchial-guided needle aspiration. However, because large-volume biopsies are needed for gene expression and tissue banking, mediastinoscopy has not lost interest, but its role is evolving. The morbidity rate associated with it is very low,[12] and the most threatening complications are haemorrhagic.

Cardiovascular diseases are the most common diseases in the general population. Recent advances in anticoagulant and antiplatelet (AP) therapies are associated with a significant risk reduction of thrombotic events. The recent recommendations of the French Societies of Cardiology and Anaesthesiology have stated that these therapies need to be conducted for several months. However, cardiovascular patients with cancer are not able to undergo surgery for exploratory or therapeutic intent because of the thrombotic risk that arises from stopping these medications prematurely. This dilemma exists because of the fear of important per- or post-operative haemorrhagic complications. Several articles have appeared recently[345] stating that general thoracic surgery is safe in patients under AP therapy. However, those papers presented various types of operations, and no specific information can be withdrawn.

The aim of this study is to report our experience with VAM in patients taking AP or anticoagulant therapies; we focus on complications (especially haemorrhagic) and how the presence of these medications can change, or not, our practice.

PATIENTS AND METHODS

From January 1, 2008, to June 30, 2012, 379 consecutive VAMs were performed in our unit. Twenty-three patients with therapeutic VAM and 34 patients who benefited in the same operative time from another diagnostic or therapeutic procedure were included in the study. To have a homogeneous group of patients, we excluded two patients with marked thrombocytopenia (platelet count <75 × 109/L) and three patients with superior vena cava syndrome from this study. Finally, 317 patients who had diagnostic VAM were reviewed for this study. Among them, 54 patients had AP (41 patients – Group A) and anticoagulant (13 patients – Group B) therapy. The control groups were formed of 263 patients (Group C).

Before surgery, the patients underwent a classic investigation workup: clinical examination, chest X-ray, CT-scan and flexible bronchoscopy for the staging of lung cancer. The blood samples were checked to have normal rates of blood cell counts and coagulation.

VAM was performed using standard equipment and following standard technique.[6] All peroperative manoeuvers needed for haemostasis (surgical clips or haemostatic sponge) were recorded. On completion of the surgery, the mediastinal dissection was not drained routinely. Frozen section analysis was done during the surgery. We used the mediastinal lymph node chart published in 1997.[7]

We have done a retrospective study from a prospectively maintained database, and we have had the Institutional Review Board approval. Data regarding the clinical records of the patients, operative time, per- and post-operative complications, total numbers of biopsies and the results of pathologic examinations were collected. In order to estimate the safety and efficacy of VAM in the presence of AP or anticoagulant therapy, we compared these two groups (A and B) of patients with the control group (C) first together (A and B vs. C), then only the AP group (A vs. C, because Group B had a low number of patients). Statistical differences between groups were calculated by the Chi-square test with P < 0.05 considered statistically significant.

RESULTS

For Group A, the indications for AP therapy were: stents in the coronary artery in 23 patients, history of coronary artery bypass and/or myocardial infarction in 15 patients, severe peripheral vascular disease in three patients. The surgery was uneventful for all patients. Two minor complications appeared. One patient experienced cardiac arrhythmia. Another patient experienced a peroperative minor haemorrhage from a bronchial artery (estimated blood loss of 100 ml); the artery was clipped. Histology samples showed lung cancer in 27 patients, other cancers in four patients and normal lymph nodes for 10 patients (VAM was done for staging purpose of lung cancer). The mean operative time was 29 min (limits 15–50 min), and the mean post-operative stay was 1.08 days (range 1–2 days).

For Group B, the indications for anticoagulant therapy were recent pulmonary embolism and/or deep venous thrombosis in seven patients, chronic atrial fibrillation in four patients and mechanical valve replacement in two patients. One minor complication appeared a peroperative minor haemorrhage from a bronchial artery (estimated blood loss of 50 ml); the artery was electrocoagulated. Histology samples showed lung cancer in seven patients and lymphoma in four patients. The mean operative time was 35 min (limits 20–50 min) and the mean post-operative stay was 1.07 days (range 1–2 days).

In the control Group C, the mean operative time was 28 min (limits 10–60 min). One death occurred (mortality rate of 0.38%) because of cardiac arrest at the induction of anaesthesia (autopsy showed myocardial infarction). One major complication occurred when one patient needed re-intubation due to severe respiratory insufficiency. Eight minor complications were noted. Two minor haemorrhages from bronchial arteries needed haemostatic clips (estimated blood loss of 100 ml and 250 ml, respectively). The other complications were 2 cardiac arrhythmia, 2 temporary recurrent nerve paralysis, 1 pleural effusion, 1 upper right lobe atelectasis, all treated effectively. The morbidity rate was 2.28%. Histology showed lung cancer in 141 patients (54%), benign disease (sarcoidosis and mycobacterial infection) in 43 patients (16%) and normal lymph nodes in 58 patients (22%). The mean post-operative stay was 1.14 days (range 1–4 days).

In Table 1, we compare the patients' characteristics concerning: age, sex, site of biopsy, total numbers of biopsies and histological results. In Table 2, we compare the results of morbidity and mortality.

Table 1

Patient’s characteristics

	Group A (n=41)	Group B (n=13)	Group C (n=263)	P A + B versus C	P A versus C
Gender (male/female)	31/10	10/3	170/93	0.5574	0.6318
Age range/median (year)	44-78/60	31-68/54	18-86/58	NS	NS
Site of biopsy (%)
2R	28 (68)	8 (61)	136 (52)	0.3475	0.3641
4R	36 (88)	11 (85)	214 (81)	0.8442	0.8526
2L	3 (7)	0	25 (9)	0.5499	0.9022
4L	8 (20)	3 (23)	51 (19)	0.9608	0.8467
7	9 (22)	3 (23)	109 (41)	0868	0.1326
Number of biopsies (total/mean)	285 (6.9)	89 (6.8)	1983 (7.5)	0.7178	0.6523
Histopathology (%)
NSCLC	24 (59)	5 (38)	122 (47)	0.6561	0.4898
SCLC	3 (7)	2 (15)	19 (7)	0.8457	0.7616
Lymphoma	2 (5)	4 (31)	16 (6)	0.3534	0.9452
Metastasis	2 (5)	1 (8)	5 (2)	0.3018	0.5556
Benign	0	0	43 (16)	0.0067	0.0207
Normal LN	10 (24)	1 (8)	58 (22)	0.9662	0.9448

NSCLC: Non-small cell lung cancer, SCLC: Small cell lung cancer, LN: Lymph node

Table 2

Operative outcome: Mortality and morbidity

	Group A (n=41)	Group B (n=13)	Group C (n=263)	P A + B versus C	P A versus C
Complication (n°/%)
Mortality	0	0	1 (0.38)	0.3784	0.2830
Minor	2 (4.87)	1 (7.69)	8 (3.04)	0.6338	0.9028
Bleeding*	1 (2.43)	1 (7.69)	2 (0.76)	0.2875	0.8797
Haemostasis (n°/%)
Clip	1	0	6	0.7510	0.6183
Sponge	3	2	15	0.5446	0.9772
Drainage	1	1	4	0.6167	0.8131
VAM mean time (min)	29 min	35 min	28 min	NS	NS
Hospital stay (days)	1.08	1.07	1.14	0.7359	0.5906

VAM: Video-assisted mediastinoscopy, NS: Not significant

DISCUSSION

Cervical mediastinoscopy is a method of exploration of the superior mediastinum. The anatomical basis is provided by the existence of cervicomediastinal continuity through specific fascia planes. From a pragmatic point of view, from cervicotomy to mediastinal dissection, there is no bleeding and 'no damageable vital structures are encountered'.[8] There are several mediastinal landmarks in the operative channel. Just below the suprasternal notch, the innominate artery crosses from the left to the right; on the left side, we have the aortic arch and the recurrent nerve; on the right side, the azygos vein forms its arch; and last but not least, the carinal region faces the pulmonary artery trunk bifurcation. All these structures are clearly identified during VAM, thus sparing us from potential complications.

Haemorrhagic complications are very rare. Any bleeding sufficient to obscure vision through the mediastinoscope is defined as major.[8] It originates from an important vascular structure and needs median sternotomy to control it. We did not experience this complication in the study period. Minor haemorrhage is related to dissection and mobilisation of the lymph nodes. Minor but important bleeding can be encountered, especially in the subcarinal nodes, from the bronchial artery that passes behind the trachea. It happened four times in our experience (1.26%). Dry gauze pressure, electrocautery or clip, usually were effective. No statistical difference was noted in our study between the groups of patients regarding these haemostatic manoeuvres, meaning that the peroperative bleeding is due to mediastinal dissection rather than AP or anticoagulant therapies. However, two tendencies were observed in our practice. The bronchial artery, when present, and of an important caliber, is now clipped instead of cauterized. When the diagnosis is obtained on the superior paratracheal nodes, and no staging purpose is needed, we do not perform a biopsy in the subcarinal nodes if the bronchial artery is present. This justifies our low rates of biopsy in the subcarinal nodes.

We thought that the presence of AP or anticoagulant therapies would change our practice. The dilemma existed if our patients could benefit from safe surgery in the presence of these medications. Data to guide the most appropriate management of patients requiring anticoagulants (AC)/AP treatment are very limited. The pre and post-operative management of those therapies is based on guidelines from the French Societies of Cardiology and Anesthesiology. For AP treatment, aspirin is maintained; clopidogrel is stopped 1 week before and reintroduced 2 days after the VAM. For anticoagulant therapy, the patients are on continuous low molecular weight heparin (that could replace the oral anticoagulant stopped 4 days earlier); the day of surgery, this treatment is stopped 12 h before, then reintroduced 12 h after the surgery.

At first, we feared the peroperative bleeding or compressive cervicomediastinal haematoma, because in our experience (before this study started), the only case of cervical haematoma, we had was on a patient taking anticoagulant therapy. The haemorrhagic complications after mediastinoscopy are very rare, mostly from large arteries or veins,[128] but like all cervical explorations, there is a potential risk for compressive haematoma.[910] We reported only one compressive haematoma at 30 days after VAM; it was related to the recurrence of cancer instead of per-operative or post-operative bleeding.[11] Moreover, we did not experience direct-related bleeding to AP/AC treatment and no cervicomediastinal haematoma. No statistical differences were noted between the groups of patients concerning the use of haemostatic manoeuvres or post-operative drainage. These manoeuvers are more related to the surgeon's appreciation of the peroperative bleeding rather than the presence of AP/AC medication. In addition, no differences were noted regarding the morbidity or mortality rates, mean operative times or post-operative stays. The only death was related to a massive myocardial infarction on a 40-year-old patient without any comorbidity in his medical history.

While the study suffers from the inherent limitations of the retrospective analysis, it represents the largest series of perioperative management of AP/AC during VAM. We tried to have homogeneous groups of patients, excluding from our study all the potential situations that can increase bleeding (abnormal coagulation or platelet count, superior vena cava syndrome, etc.). Perhaps, there are limitations in our study, first, because all the patients had AP treatment exclusively with aspirin (except one patient under clopidogrel). The platelet inhibition mechanisms are different. Clopidogrel acts by inhibiting adenosine diphosphate-dependent platelet activation, and aspirin acts by inhibiting thromboxane-dependent platelet activation.[4] Second, we mixed anticoagulant and AP therapy groups of patients to have a considerable number of patients to study. However, we compared separately the AP group versus the control group; no statistically significant differences were noted. Another bias is the ability to match the control group. In general, all the patients of the AP/AC groups had lung or mediastinal malignancies, while the control group had a noteworthy percentage of benign pathologies.

CONCLUSION

We have shown that the VAM can be safely performed in patients receiving AP or anticoagulant treatments. There is no increase of peroperative bleeding or post-operative compressive cervicomediastinal haematoma. The AP or anticoagulant therapies must be continued to prevent catastrophic thrombotic events.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.