HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Muir, A.D. Articles by Graham, A.N. Search for Related Content PubMed PubMed Citation Articles by Muir, A.D. Articles by Graham, A.N. Related Collections Esophagus - other

Eur J Cardiothorac Surg 2003;23:799-804
© 2003 Elsevier Science NL

Treatment and outcomes of oesophageal perforation in a tertiary referral centre

Department of Thoracic Surgery, Royal Victoria Hospital, Grosvenor Road, Belfast, BT12 6BA, Northern Ireland, UK

Received 3 September 2002; received in revised form 19 December 2002; accepted 22 January 2003.

^* Corresponding author. Tel.: +44-2890-240503; fax: +44-2890-314159
e-mail: alastair.graham{at}royalhospitals.n-i.nhs.uk

	Abstract

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion Appendix A. Conference... References

 
Objective: The diagnosis and management of oesophageal perforation continues to challenge clinicians. We present our experience of perforated oesophagus in a Tertiary Referral Centre for Thoracic and Oesophageal Surgery. Methods: Between 1985 and 2000, 75 patients (40 male) with oesophageal perforation were treated in out unit; age range 24–89, median 63. Retrospective review of these cases has been performed. Results: There were 12 deaths (16%). With increases in time from perforation to diagnosis, there was a stepwise increase in the mortality rate. Immediate diagnosis 5%; early diagnosis (1–24 h) 14%; late diagnosis (>24 h) 44% (P0.002). Site of perforation, aetiology, and treatment strategy had no influence on mortality. The only independent predictor of mortality identified was time to diagnosis from perforation (beta 0.429, P=0.001). Time to definitive management in those undergoing an operative procedure had no influence on outcome with multivariate analysis. Conclusions: Prompt recognition of the diagnosis of oesophageal perforation and rapid institution of supportive measures, followed by an appropriate, patient specific treatment option optimises the chance of a successful outcome. The wide range of presentation of oesophageal perforation necessitates individualisation of treatment.

Key Words: Oesophageal perforation

	1. Introduction

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion Appendix A. Conference... References

 
Oesophageal perforations have long been recognised to be a cause for both high morbidity and mortality rates. Despite continuing advances in intensive patient support, published mortality ranges from 9 to 41% [1–6]. As the regional Thoracic surgical unit, we provide a tertiary referral service for all oesophageal perforations occurring in Northern Ireland, population 1.7 million, and therefore have a heterogeneous group of patients presenting for treatment. We are involved with the care of all patients diagnosed with oesophageal perforation at our institution, and believe that very few patients from other hospitals in our region are managed without our advice being sought. We report upon our experience of oesophageal perforations seen over a 16-year period (January 1985–December 2000).

	2. Patients and methods

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion Appendix A. Conference... References

 
Patients were identified from hospital records, operating theatre logbooks, and in the last 6 years, from the computerised thoracic surgical database. In total, there were 75 patients with the diagnosis of oesophageal perforation; (40 male (53%)). Their mean age was 61.3 (SD 15.3, range 23–89) years. The diagnosis of oesophageal perforation was confirmed by oesophagoscopy, radiological contrast swallow study, by the presence of mediastinal air or subcutaneous empysema.

2.1. Statistical analysis
Univariate analysis methods included Student's t-, Fisher's exact and ² tests. The relationships between multiple factors were examined using linear regression analysis. Statistical analysis was performed using STATISTICA v5.0 for Windows (StatSoft Inc., Tulsa Oklahoma). A P value of less than 0.05 was considered to be significant.

	3. Results

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion Appendix A. Conference... References

 
Seventy-five patients presented to our service between January 1985 and December 2000 with a diagnosis of oesophageal perforation. There were 12 deaths in total, giving an overall mortality of 16%.

3.1. Aetiology
Fifty-six of 75 patients (75%) presented following instrumentation of the oesophagus, 13 had a spontaneous cause for their perforation, with the remaining six patients having other aetiologies (Table 1). The mortality for perforation due to instrumentation of the oesophagus was 14%, with eight deaths among 56 patients. Spontaneous perforation of the oesophagus resulted in a mortality of 15%, three deaths from 13 patients. Those patients who presented with a classical Boerhaave's syndrome history had a mortality of 10% (P>0.05 compared with instrumental perforation).

3.2. Site
The majority of perforations were situated in the thoracic oesophagus (Table 2). Perforation of the cervical oesophagus was associated with a lower mortality than thoracic perforation, but this failed to reach statistical significance.

View this table: [in this window] [in a new window]   Table 3. Influence of time from perforation to diagnosis on mortality (Pearson's 2 12.349 (two degrees of freedom) P=0.002)

3.5. Treatment methods
Overall, 58 patients underwent a surgical intervention (77%), and 17 were conservatively managed (23%) (Table 4). During the interval this study covers, the philosophy of treatment for perforations of the oesophagus was to excise any perforation of a resectable tumour, and to treat non-malignant perforations with less extensive surgical interventions.

3.7. Drainage
The largest group of patients underwent an operative drainage procedure of some form. The aim of these procedures was drainage of any intra-thoracic collection, to minimise further loss of gastrointestinal contents, and to provide a route for enteral feeding (typically by insertion of jejunostomy). Three patients in this group had oesophageal cancer, but had previously been assessed and felt unfit for oesophageal resection.

Of these 30 patients, 20 presented within 24 h of their perforation, with only one death (5%), whilst four of the ten presenting late died (40%) (P<0.05).

3.8. Resection
Of the 15 patients who underwent a resection, there were two deaths. Of the 15 patients undergoing resection of their oesophagus, ten had oesophageal cancer. The other five patients had resections for a variety of other reasons. Two were for benign strictures in the presence of oesophageal metaplasia (Barrett's oesophagus), and the others for one each of achalasia, oesophageal varices, and following perforation after repair of a para-oesophageal hernia.

In this group, two treatment strategies were employed. Resection and reconstruction of the perforated oesophagus at initial operation, or resection alone, with delayed reconstruction when the patient's clinical condition allowed. Of the 13 survivors, two had staged procedures. Of the patients who died, one had reconstruction at the initial operation, and the other had resection alone, due to his unstable clinical condition. All of the resections were performed on patients who were diagnosed within 24 h.

3.9. Primary repair
Of the 13 patients who had their perforation repaired without resection, only one died (8%). No patient undergoing primary repair of his or her oesophagus had malignancy. Three differing strategies were used. Nine patients had oversewing of the perforation alone. Two patients had repair of the perforation, with an anti-reflux procedure also (Nissen's and Belsey wraps), whilst two had repair, wrap and a contralateral myotomy of the lower oesophagus (Heller's type). The death in this group was a patient who underwent repair with Nissen's fundoplication 6 h following instrumental perforation. All of the primary repairs were performed on patients who were diagnosed within 24 h.

3.10. Conservative
Of the 17 non-operatively managed patients, four died (24%). Of these patients, 12 had benign disease, three had oesophageal tumour, but were deemed unfit for a surgical procedure, and two had malignancy remote from their oesophagus (breast and pancreatic tumours). Management strategies included additional oxygen, intravenous fluids and antibiotics, insertion of nasogastric tube for decompression of the stomach (performed under radiological guidance), nutritional support (enteral and/or parenteral), and insertion of appropriate percutaneous drains. In comparison there was 14% mortality in the surgically managed group. (P>0.05). Eleven non-operatively managed patients presented within 24 h, with one death, compared to three deaths of six presenting late (P>0.05).

3.11. Influence of malignancy
A total of 20 patients had an underlying diagnosis of malignancy. One of these patients had pancreatic carcinoma, and her oesophageal perforation occurred during ERCP and stent insertion for common bile duct stenosis. One patient had perforation during dilatation of a chemotherapy induced oesophageal stricture. Seventeen patients had oesophageal cancer. In this sub-group of patients, there were four deaths (P>0.05 compared to benign disease).

3.12. Influence of pre-operative co-morbidity
Of the 12 patients who died, six had pre-existing cardio-respiratory disease (history of myocardial infarction, angina pectoris, COPD, atrial fibrillation, chronic heavy cigarette smoking). Of the patients who survived, 23 of 63 had similar pre-existing disease (P>0.05).

3.13. Multivariate analysis
Multivariate analysis was performed looking at the dependent variables death, hospital stay, intensive care stay (>5 days), and the independent variables gender, age, aetiology, site of perforation, extent of perforation, pre-operative cardio-respiratory disease, time to suspicion of diagnosis, time to procedure from presentation, year of presentation and specifically if the aetiology was spontaneous or instrumental.

Of all variables, only time from perforation to suspicion of diagnosis (with initiation of initial management) was predictive of mortality (P=0.001, beta=0.429 backward stepwise).

Although univariate analysis shows an increasing mortality rate with increasing time from perforation to diagnosis and from diagnosis to eventual surgical intervention (Table 5), in multivariate analysis there was no factor that independently predicted outcome.

	4. Discussion

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion Appendix A. Conference... References

Pre-existing cardio-respiratory disease did not influence the eventual outcome. However, roughly half of those who died had a history of some form of cardio-respiratory disease, whereas roughly one third of those who survived have a similar disease profile. With larger numbers, these differences may well have been significant. It has been our practice to treat perforated oesophageal cancer with resection, and to tend towards conservation of the oesophagus with benign disease.

In our institution, we are consulted on all perforations of the oesophagus, regardless of the site or cause of the perforation. We use wide-ranging methods for treating our patients, from simple supportive measures, through oesophageal stenting and drainage procedures to radical excisions of the intra-thoracic oesophagus, with outcomes comparable to other published series [1,15,16].

We chose two specific time periods during the course of a patient's treatment for analysis. The first was the length of time from onset of symptoms (the actual perforation of the oesophagus) to diagnosis of that perforation. The second was the length of time from diagnosis of oesophageal perforation to surgical procedure. The time from perforation to diagnosis was chosen because it is normally an exactly identifiable time period, and is not influenced by subsequent treatment. The second time-period was chosen because it does not overlap the first time-period, and is a measure of the patient's treatment in our institution, whatever the modality of treatment.

Analysing our outcomes, the most influential factor is the speed with which the perforation is recognised and appropriate treatment instigated. There were no demonstrable differences in outcome between treatment modalities. Oesophageal perforation is a rare condition, so despite this being one of the largest reported series, numbers in each group for comparison are small. There are inherent pitfalls when conclusions are drawn from detailed analysis of small numbers of patients. There may very well be differences between survivors and those who die, but the analysis lacks the required power to discriminate.

Univariate analysis indicates that both time from perforation to diagnosis, and time from diagnosis to surgical intervention have influence upon outcome. With multivariate analysis, however, time from diagnosis to surgical intervention fails to reach significance, whilst time from perforation to diagnosis is strongly predictive of outcome (P=0.001). This point is notably different from previously published series. This, taken together with the fact that conservatively managed patients do no worse than those receiving surgery (P>0.05), leads us to believe that the initial steps taken upon admission with perforated oesophagus, namely resuscitation, nil orally and intra-venous broad spectrum antibiotics are of utmost importance. These steps will limit further loss of gut contents from the oesophagus, whilst supporting the circulation, and combating infection. In patients proceeding to surgery, this amounts to pre-operative optimisation, a step which leads to improved outcomes in any circumstance.

We feel that these findings support the view that patients with oesophageal perforation should not automatically be taken to an operating theatre upon their diagnosis. Instead, a planned intervention undertaken on the next available operating list may be more appropriate, allowing an adequate period of resuscitation for the patient, and meaning that a fresh, complete operating team can perform what can be demanding surgery [17].

Historically, spontaneous rupture of the oesophagus, namely Boerhaave's syndrome has been associated with a high mortality. This is not the case with our experience. Indeed, those patients with a classical Boerhaave's syndrome presentation had a mortality of 10%, with the mortality of patients with instrumental perforation being 14%. This runs contrary to a widely held belief that spontaneous perforations should fare badly compared to instrumental perforations.

Compared to older publications [5,6,11], the mortality in our series is favourable, illustrating the increasing effectiveness of management strategies and the advances in care for the critically ill patient.

The lowest mortality rate was seen in the group of patients who had a cervical perforation (8%). However, this difference was not statistically significant from those with other sites of perforation. Lack of spillage of gut contents into the pleural spaces or peritoneum may account for this tendency towards survival, however, it must be remembered that infection tracking along tissue planes into the mediastinum is still a danger.

With perforation of the thoracic oesophagus, spillage of gut contents into the pleural cavities had a higher (22 vs. 11%) mortality than when soiling was confined to the mediastinum (P>0.05). There were greater numbers of patients with pleural soiling presenting after 24 h. With multivariate analysis, whether there was spillage into the pleural cavities did not influence outcome.

This study is a retrospective review of our practice over the last 16 years, and so the interpretation of the data is more difficult than with other types of study design. There is no consistent objective physiological assessment available for comparison of these patients, clouding the issue further. Despite these shortcomings, this investigation is the third largest in the literature concerning oesophageal perforation, so we feel that we can make some recommendations on management and treatment options for this condition.

1. The speed with which the correct diagnosis is made, and appropriate therapy commenced is paramount.
   
2. The extent and severity of perforation will have a large influence on chosen treatment option.
   
3. Patient fitness and any pre-existing oesophageal disease will help determine the preferred management option.
   
4. Surgical intervention, if felt to be the best option, should be performed at the start of the next working day, following the commencement of initial conservative measures and resuscitation.
   

In conclusion, oesophageal perforation will continue to be a challenging, if rare, condition to manage effectively. Mortality rates will continue to frustrate, although there may be further reductions in the overall mortality rate. With the increasing incidence of oesophageal instrumentation (both diagnostic and therapeutic), the incidence of perforation will probably not diminish.

	Acknowledgments

 
We would like to thank Dr C. Patterson, PhD for his assistance with the statistical analysis.

	Footnotes

 
Presented at the 16th Annual Meeting of the European Association for Cardio-thoracic Surgery, Monte Carlo, Monaco, September 22–25, 2002.

	Appendix A. Conference discussion

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion Appendix A. Conference... References

 
Dr T. Dosios (Athens, Greece): There are two questions and one comment. First question: how do you select patients for conservative or surgical treatment? Second question: what are your criteria to operate on a patient who is already in conservative management? The comment is: I was impressed that very few of your patients were attempted to have a primary repair; it was only about 25%. Can you comment on that, please?

Dr Muir: The treatment modalities are based entirely upon the patient's clinical condition. We feel that each patient is an individual and presents in an individual manner after a particular form of insult to their esophagus. Patients are selected for various treatment methods on a day-to-day basis. If someone has been selected for conservative management and things are starting to go wrong, some of these patients do eventually progress to a surgical procedure, but that is at an at-ward level basis depending on a multitude of factors. We do not have any strict criteria that I could lay out for you.

Dr N. Metity (Baghdad, Iraq): It is rare to see a conservative treatment that is helpful in the intrathoracic perforation. You have shown that by draining the mediastinum, or whatever, you got very good results. I wonder what is the percentage of success in the intrathoracic perforation.

Dr Muir: One of the slides showed that 80% of our patients have their perforation in the intrathoracic esophagus, and the outcome for that group of patients is 17%. This is in comparison with the cervical esophagus being 9% and the abdominal esophagus being around 33%. So these patients do equally well. The subgroup of patients who are conservatively managed are patients from all three groups, and there are no differences in outcome. We based our treatment methods, as I said to the previous question, on the clinical condition of each patient.

I think this is a very provocative paper, to say the least, because in one of your earlier slides you say that conservative treatment had a much higher mortality than operative intervention. I think what is important in your message is that really it is the individualization of the treatment that is important in deciding what to do, and also for the patient to be seen by an esophageal expert.

How do you account for the delay in a lot of your patients? I think, as Dr. Dosios has already pointed to, how do you decide which is the best therapy. I think the decision-making process is extremely difficult.

Have you any radiographic criteria, for instance, if the mucosa has been breached or there is a pleural effusion or age? I agree with you that you should resuscitate first, but how do you decide whether to operate on these patients?

Dr Muir: I work in a regional unit. We have patients referred to us from all other regions with a population of around 2 million. If we receive the patients and we understand that they have an esophageal perforation straight away, the patients do very well. Many of the patients have a delayed diagnosis; have had their diagnosis made at another institution and transferred to ourselves.

Sorry, the second part of your question?

Mr Thorpe: I was really wanting to get a feeling of what are the causes of delay and what are your main criteria for really operating on someone.

Dr Muir: Basically it is built upon a clinical condition, what the patient's preoperative diagnosis was. For instance, I said there were a number of patients that had esophageal tumors. Those patients all had operative management because there is no way that you can conservatively manage and then go on to further treat their esophageal cancer. You are creating problems for yourself. Really, we don't have such a criteria.

But the initial belief in the regional unit in the past was an esophageal perforation should equal operation, and due to these figures, in their preparation in the previous months and years, a shift in thinking has occurred among the senior surgeons away from immediate surgery. And that is the message I am trying to convey. The fact is these patients are a sick group of patients and will need treated appropriately. We are just calling into question the fact that rushing headlong to theater may not be the best thing for them.

	References

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion Appendix A. Conference... References

 

1. Reeder L.B., DeFilippi V.J., Ferguson M.K. Current results of therapy for oesophageal perforation. Am J Surg 1995;169:615-617.[Medline]
2. Iannettoni M.D., Vlessis A.A., Whyte R.I., Orringer M.B. Functional outcome after surgical treatment of oesophageal perforation. Ann Thorac Surg 1997;64:1606-1610.[Abstract/Free Full Text]
3. Kotsis L., Kostic S., Zubovits K. Multimodality treatment of oesophageal disruptions. Chest 1997;112:1304-1309.[Abstract/Free Full Text]
4. Okten I., Cangir A.K., Ozdemir N., Kavukcu S., Akay H., Yavuzer S. Management of oesophageal perforation. Surg Today 2001;31(1):36-39.[Medline]
5. Wilde P.H., Mullany C.J. Oesophageal perforation – a review of 37 cases. Aust N Z J Surg 1987;57:743-747.[Medline]
6. Pate J.W., Walker W.A., Cole F.H., Owen E.W., Johnson W.H. Spontaneous rupture of the oesophagus: a 30-year experience. Ann Thorac Surg 1989;47:689-692.[Abstract]
7. Mamun M. Spontaneous oesophageal perforation: long known but still not easy to diagnose. Hosp Med 1998;59(12):968-969.[Medline]
8. Henderson J.A., Peloquin A.J. Boerhaave revisited: spontaneous oesophageal perforation as a diagnostic masquerader. Am J Med 1989;86(5):559-567.[CrossRef][Medline]
9. Scheinin S.A., Wells P.R. Oesophageal perforation in a sword swallower. Tex Heart Inst J 2001;28(1):65-68.[Medline]
10. Jones W.G., Ginsberg R.J. Oesophageal perforation: a continuing challenge. Ann Thorac Surg 1992;53(3):534-543.[Abstract]
11. Pillay S.P., Ward M., Cowen A., Pollard E. Oesophageal ruptures and perforations – a review. Med J Aust 1989;150:246-252.[Medline]
12. Altorjay A., Kiss J., Voros A., Bohak A. Nonoperative management of oesophageal perforations. Is it justified?. Ann Surg 1997;225(4):415-421.[CrossRef][Medline]
13. Altorjay A., Kiss J., Voros A., Sziranyi E. The role of oesophagectomy in the management of oesophageal perforations. Ann Thorac Surg 1998;65:1433-1436.[Abstract/Free Full Text]
14. Bufkin B.L., Miller J.I., Mansour K.A. Oesophageal perforation: emphasis on management. Ann Thorac Surg 1996;61:1447-1452.[Abstract/Free Full Text]
15. Lawrence D.R., Moxon R.E., Fountain S.W., Ohri S.K., Townsend E.R. Iatrogenic oesophageal perforations: a clinical review. Ann R Coll Surg Engl 1998;80:115-118.[Medline]
16. Fernandez F.F., Richter A., Freudenberg S., Wendl K., Manegold B.C. Treatment of endoscopic oesophageal perforation. Surg Endosc 1999;13:962-966.[CrossRef][Medline]
17. NCEPOD who operates when? 1995/96 Report. http://www.ncepod.org.uk/reports.htm.

This article has been cited by other articles:

				I. Paul, B. Badmanaban, and A.N.J. Graham Perforation of the lower thoracic oesophagus following crush injury to the chest and abdomen Eur. J. Cardiothorac. Surg., March 1, 2005; 27(3): 526 - 528. [Abstract] [Full Text] [PDF]

				A. D. Muir and A. Graham Primary esophageal repair for Boerhaave's syndrome whatever the free interval between perforation and treatment Eur. J. Cardiothorac. Surg., February 1, 2005; 27(2): 356 - 356. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Muir, A.D. Articles by Graham, A.N. Search for Related Content PubMed PubMed Citation Articles by Muir, A.D. Articles by Graham, A.N. Related Collections Esophagus - other