Diagnosis, management and outcomes of thoracic esophageal perforation

Introduction

Esophageal perforation has been considered a catastrophic and often life-threatening event, with very high morbidity and greater than 20% mortality rates even though appropriate treatment is started on time. About half of esophageal perforations are iatrogenic, mostly as a result of endoscopic instrumentation maneuvers, and more then a third are spontaneous rupture of oesophagus associated with forceful vomiting after overconsumption in food and alcohol which is known as Boerhaave’s syndrome. Delay in diagnostics and carrying out of surgical intervention, that occurs in more than 50% of cases, is a major factor of high mortality rate which rises from 40% to 60% if diagnosis and treatment are delayed, but this rate decreases to 10–25% if treatment is carried out within 24 h of perforation [1].

Aim of the research

Our aim was to present the results and difficulties in the diagnosis and management of esophageal perforation based on the experience of our department of thoracic reconstructive surgery as well as data obtained from 6 other hospitals.

Material and methods

We performed a retrospective analysis of the management of 103 patients (mean age: 49.4 ±3.1 years; 18–82, 80 males and 23 females) treated for thoracic esophageal perforation at our hospital (16 patients) and the other 6 hospitals of (87 cases) during the period of 1997–2011. Nineteen (18.4%) patients (mean age: 64.7 ±5.7) had comorbidities with a prevalence of the cardiovascular system pathology – in 11 (57.8%), a respiratory disease encountered in 4 (21.1%), diabetes in 3 cases (15.8%) and the digestive system pathology – in 1 (5.3%).
All patients were divided into two groups for analysis of different surgical techniques: the first group (historical control group) – 94 (91.3%) patients who were treated by open thoracotomy or/and laparotomy (“open surgery” – OS) – was compared with the second “video-assisted thoracoscopic surgery (VATS)group” of 9 patients who had undergone VATS non-resection drainage procedure in our hospital.
Chest computed tomography with oral contrast was used in all patients of “VATS group” to confirm oesophageal perforation and its healing after surgery. Videothoracoscopy was carried out through the left pleural cavity for patients with injury in the lowest third thoracic esophageal (n = 3) and through the right pleural cavity in case of perforation of a medial part (n = 6). Application of a primary repair with a suture of esophagus wound was not indicated because of delay in the surgery (88.1 ±61.7 h).
Perforation of a low third thoracic esophagus was accompanied by a fistula to the left pleural cavity and pyopneumothorax. At the moment of operation these patients had the expressed clinical symptoms of purulent fever and severe cardiac comorbidity. In this situation it was decided to install drainage for the sanity of any pleural purulent collections. The access to esophagus was carried out through the left pleural cavity by insertion of optics and tools through of the three chest ports (Figure 1). The massive purulent peel on the pleural surface and exudate were determined during an observation of pleural cavities. Endoscopy examination during operation allowed to find the localization of esophageal defect. With the use of the technique developed by us [2], the polyvinyl drainage tube was inserted through a pleural cavity into esophageal perforation and then fixed by means of a nasojejunal tube in esophageal lumen (Figure 2). This method of «fistulazation» allowed to avoid the formation of gastrostoma for feeding and to perfom adequate drainage of pus pleural effusion from the area of injury and the pleural cavity. Also, three chest ports accesses were carried out through the right pleural cavity for the management of patients with median third of the esophageal perforation. The post-operative feeding through the nasojejunal tube began on the first day after operation as well as early physical activity of patients. The intrathoracic fistula drainage tube was withdrawn little by little each day as soon as no evidence of leakage into the thoracic cavity was observed by esophagogram.
The open surgical management of esophageal perforation in the control group included a primary repair with (20 cases) or without cervical oesophagostomy and feeding gastrostomy in 54 (57.5%) patients, a thoractomy or laparotomy or cervicotomy to drain and debride the pleural and mediastinal collections with (14 cases) or without oesophageal diversion and gastrostomy in 32 (33.9%), an esophageal resection with cervical oesophagostomy and gastrostomy in 5 (5.3%) and thoracotomy with modified by us procedure of the T-tube drainage for a “controlled esophago-cutaneous fistula” in 3 cases (3.3%).

Statistical analysis

Pseudorandomization was made to avoid heterogeneity in the comparison of groups by gender, age, clinical diagnosis and time from the injury of the esophagus until the surgical care. There were no significant differences in age and gender, the structure of diagnoses and time interval after the perforation (Table 1) or presence of comorbidity between the two groups.
The descriptive statistics were summarized as a means for continuous variables (± standard error from the mean) and as frequencies and percentages for categorical variables on Microsoft Excel for Window. The Mann-Whitney U test was performed to calculate differences in continuous variables and differences between categorical variables were tested with Fisher’s exact test. All differences were considered significant at a p value less than 0.05.

Results

The most common localization of thoracic esophageal perforation was the distal third (n = 58, 56.4%), followed by the middle third (n = 28, 27.2) and the proximal third (n = 17, 16.4%).
Most of the injuries of the thoracic esophagus were of iatrogenic nature – 42 (40.7%), due to perforation occurred during the post-burn scar stricture bougienage – 29 patients (28.2%), fibroesophagoscopy patients – 11 (10.6%) and rigid esophagoscopy – 2 (1.9%), respectively. Boerhaave’s Syndrome was detected in 39 cases (38%) and the esophageal foreign body trauma in 16 (15.5%). The remaining 6 cases (5.8%) occurred as a result of the external traumatic agent: 3 patients (2.9%) – a chemical burn of the esophageal wall, in 2 (1.9%) – stabbed and 1 (1%) patient as a result of the explosive injury. There were no significant differences in the etiology of perforation between the two groups.
Eighty-four patients (81.6%) out of 103 were delivered in the first 24 h to the hospital Emergency Department due to the severity of the disease symptoms that made patients to seek medical help within an average pre-hospital period of 20.6 ±9.6 h.
Despite the fact that more than 80% of patients were hospitalized on the first day, in 42 cases (40.8%), surgical treatment was applied after 24 h with the mean time interval between perforation and operation of 52.1 ±7.8 h (Table 2).
Data analysis has revealed that 32 (31%) of all patients were not recognized as a “thoracic esophageal injury” during the first examination. The most common initial “mistakes” were perforation of duodenum ulcer and polysegmental pneumonia on 8 cases for each pathology (25%). Four patients (12.5%) were treated on suspicion of pancreatitis, pleurisy or pneumohydrothorax – 3 (9.5%), and as the beginning of neck phlegmonous adenitis – 2 cases (6.3%). Moreover, one case of paratonsillitis, myocardial infarction, gastric bleeding, decompensated pyloric stenosis, caustic esophageal burn, an esophageal foreign body without perforation and acute hepatitis have been wrongly diagnosed at the first examination. These diagnostic mistakes led to unnecessary surgery in 13 patients such as diagnostic laparotomy – 10 cases, drainage of the pleural cavity in 2 and the resection of the stomach in one case.
An initial chest X-ray in the emergency department let to raising suspicion for oesophageal perforation in 55 out of 78 patients (68%) so as most X-rays had more than one sign of abnormality: unilateral pleural effusion and/or pneumothorax (Figure 3), pneumomediastinum, subcutaneous emphysema (Figure 4). In the remaining 23 patients, however, chest X-rays didn’t show any abnormalities or radiological findings were misidentified as infiltrative changes in the lungs.
Eighty-one (78.6%) out of the 103 patients under­went oesophagography with water-soluble contrast to confirm perforations, 66 (81.5%) of which defini­tely confirmed the pathology showing contrast leakage through esophageal wall into pleural cavity and pneumo- and hydrothorax. Ten foreign body perforation patients and five iatrogenic etiology (3 – bougie, 2 – endoscope) had false negative oesophagographies. Furthermore, an incorrect interpretation of the patient condition and the choice of inadequate treatment were undertaken in 3 (3.7%) patients with negative imaging results. Thirteen patients (12.6%) underwent chest computed tomography (CT) scans with 100% diagnostic efficiency in patients with atypical presentations by means of imaging direct radiological evidence of esophageal perforation such as esophageal wall discontinuity (Figure 5) and a contrast leak (Figure 6) or/and indirect signs which include pleural effusion, pneumomediastinum (Figure 7), subcutaneous emphysema, hydrothorax, pneumothorax and collapse of the lung (Figure 8).
Forty-one patients (38.9%) required a flexible oesophagoscopy to confirm the perforation because the chest X-ray and contrast swallow results were equivocal, or when CT scans with orally administered contrast medium were impossible for technical reasons. In all cases during diagnostic endoscopy the defect of esophageal mucous membrane was revealed, and two of them had a fistula to the pleural cavity.
Sixty-five (69.2%) patients of the control group were complicated by postoperative morbidity that included postoperative leakage, pleural empyema, pneumonia, pulmonary embolism, pericardial effusion, gastric bleeding, a Esophagus fistula with osteomyelitis of the rib and wound infections that resulted in high mortality rate (33 cases). The median hospital stay to discharge or death was 50.9 ±8.7 among those patients who had postoperative morbidity whereas without these complications it was 29.7 ±4.44 days.
Postoperative leakage occurred in 32 patients among 54 (62.3%) who underwent primary repair of the thoracic esophagus perforations, more over the primary repair delayed for more than 24 h resulted in doubling of the postoperative leakage rate (22 cases).
For the VATS group of patients, the postoperative morbidity and mortality rates were lower (11.1% and 11.1%, p < 0.05), as only one patient with delayed diagnosis of Boerhaave’s syndrome, died on the 28th postoperative day of respiratory failure, which was caused by a postoperative hospital-acquired pneumonia, mediastinitis and systemic sepsis (Figure 9).
The median time in surgery and the in-hospital stay also was shorter (p < 0.05) in the VATS-surgery group (64.2 min and 26.5 ±5.6 days) than in the open-surgery group (116.6 and 41.2 ±6.1 respectively) (Table 3).

Discussion

Esophageal perforations, particularly spontaneous, are difficult to diagnose because they can masquerade many clinical conditions like acute myocardial infarction, acute aortic dissection, tension pneumothorax [3], pneumoperitoneum [4] or different digestive system pathology as it has been presented in our study.
That is why the diagnosis can easily be missed or delayed, leading to delaying in the treatment in more than 50% of cases which accounts for the high mortality rate of 20–75% or 100% if left untreated [5].
As can be seen from our series, the delayed surgery in 42 patients (41%) is fraught high postoperative morbidity (66 patients, 64%) and 52% (n = 22 out of 42) in-hospital mortality, whereas overall mortality rate was 33% (n = 34 out of 104).
Several studies have also found that late (> 24 h) diagnosis and diagnosis of perforations may take place at centers away from the ones being appropriately managed are the main factors associated with poor outcomes [6, 7].
The most common misdiagnosis is perforated ulcer, myocardial infarction, pulmonary embolism, dissecting aneurysm and pancreatitis [8, 9]. In our data collective, thirteen patients were undergone unnecessary exploration surgery because of initial diagnostic mistakes and most often it was done in hospitals with inadequate clinical experience in thoracic surgery. The most common reasons for a delay in diagnosis were initial misdiagnosis and in 2 cases there were late presentations of patients.
We support the opinion of other researchers that diagnosis can be made earlier and more accurate with CT scan [10, 11], because the use of CT with contrast may expedit the diagnosis even in critically ill patients with atypical symptoms which might impede rapid identification of esophageal perforation. More over we consider, that up today CT scans may be useful not only for study of fluid and air collections, the extent of a surrounding inflammatory process but it can precisely enough localize the site of the perforation before VATS surgery and illustrate the postoperative healing of a esophageal perforation. Thus, from our point of view, CT with water-soluble contrast may become the “gold standard” in the diagnosis and management of esophageal perforation.
Despite this the endoscopy has sensitivity of 100% with specificity of 83% [12]. we recommend performing upper endoscopy only if perforation is not found with the contrast X-ray and/or CT, because it can increase the size of the already existing perforation and also introduce more air into the mediastinum.
We agree with many authors that for detected early perforations (less than 24 h from injury), the best treatment is primary repair [13–15].
However, we disagree with the view of Jougon et al. [16] that primary repair may have satisfactory results regardless of the time interval from injury to repair.
Our experience is that nobody of the VATS-group patients was fit for attempted primary surgical repair because the esophageal wound edges were oedematous, friable and also there was often associated mediastinitis and/or empyema rendering primary repair too risky. Depending on the localizations of thoracic esophageal perforations we decided to use either a left – or right-sided VATS approach for adequate surgical debridement and drainage of the mediastinum by the like T-tube drainage procedure. Our experience with VATS forming esophagocutaneous fistula shows that this approach could be used as the first choice for management of esophageal perforations regardless of etiology and it is effective by minimizing additional surgical trauma in these high morbidity patients. More over, the survival results were much better than those for open surgery, however, less than hundred percent as Vogel et al. have published [17].

Conclusions

Esophageal perforations is a rare pathology and due to the rarity of this condition and its often nonspecific presentation, the surgical treatment of this is delayed in more than 40% of patients, which leads to a death of every third patient.
Computer tomography with the oral contrast swallow must be the first-line investigation of patients who present signs or are suspected of esophageal perforation to speed up diagnosis and get confirmation of the site and extent of perforation in order to expedit the choice of an optimal treatment.
The developed by us video-assisted thoracoscopy method of treatment of the delayed (more than 24 h) esophageal perforation possesses a smaller operation trauma in comparison with the standard open surgical techniques which allows to apply this method to the treatment of patients in bad condition with an accompanying pathology and with severe inflammatory reaction.

References

1. Bladergroen MR, Lowe JE, Postlethwait RW. Diagnosis and recommended management of esophageal perforation and rupture. Ann Thorac Surg 1986; 42: 235-9. 
2. Panko S, Karpitski A, Shestiuk A, et al. The advantages of videoassisted treatment of delayed cases of esophageal perforation. Studia Medyczne 2008; 12: 7-12.
3. Onyeka WO, Booth SJ. Boerhaave’s syndrome presenting as tension pneumothorax. J Accid Emerg Med 1999; 16: 235-6.
4. Witz M, Jedeikin R, Zager M, et al. Spontaneous rupture of the distal oesophagus (Boerhaave’s syndrome) with unusual clinical presentation of pneumoperitoneum. Postgrad Med J 1984; 60: 60-1.
5. Turut H, Gulhan E, Adams PY, et al. Successful conservative management of Boerhaave’s syndrome with late presentation. J Natl Med Assoc 2006; 98: 1857-9.
6. Muir AD, White J, McGuigan JA, et al. Treatment and outcomes of oesophageal perforation in a tertiary referral centre. Eur J Cardiothorac Surg 2003; 23: 799-804.
7. Reeder LB, DeFilippi VJ, Ferguson MK. Current results of therapy for esophageal perforation. Am J Surg 1995; 169: 615-7.
8. Symbas PN, Hatcher CR, Harlaftis N. Spontaneous rupture of the oesophagus. Ann Surg 1978; 187: 634-9.
9. Curcij JJ, Horman MJ. Boerhaave’s syndrome: the importance of early diagnosis and treatment. Ann Surg 1976; 183: 401-8.
10. De Lutio Di Castelguidone E, Pinto A, et al. Role of spiral and multislice computed tomography in the evaluation of traumatic and spontaneous oesophageal perforation. Our experience. Radiol Med (Torino) 2005; 109: 252-9.
11. Landen S, El Nakadi I. Minimally invasive approach to Boerhaave’s syndrome: a pilot study of three cases. Surg Endosc 2002; 16: 1354-7.
12. Horwitz B, Krevsky B, Buckman RF. Endoscopic evaluation of penetrating esophageal injuries. Am J Gastroenterol 1993; 88: 1249-53.
13. Port JL, Kent MS, Korst RJ, et al. Thoracic esophageal perforations: a decade of experience. Ann Thorac Surg 2003; 75: 1071-4.
14. Gupta NM, Kaman L. Personal management of 57 consecutive patients with esophageal perforation. Am J Surg  2004; 187: 58-63. 
15. Andrade-Alegre R. Surgical treatment of traumatic esophageal perforations: analysis of 10 cases. Clinics (Sao Paulo) 2005; 60: 375-80. 
16. Jougon J, McBride T, Delcambre F, et al. Primary esophageal repair for Boerhaave’s syndrome whatever the free interval between perforation and treatment. Eur J Cardiothorac Surg 2004; 25: 475-89. 
17. Vogel SB, Rout WR, Martin TD, et al. Esophageal perforation in adults: aggressive, conservative treatment lowers morbidity and mortality. Ann Surg 2005; 241: 1016-23.

Address for correspondence:

Prof. Siarhei Panko MD, PhD
Department of Emergency Medicine
Faculty of Health Sciences
Jan Kochanowski University
al. IX Wieków Kielc 19, 25-317 Kielce, Poland
Phone: +375291108819
E-mail: sppankam@gmail.com, pan@brsu.brest.by