Eur J Cardio-thorac

Surg (1992) 6:407-411

0 Springer-Verlag 1992

Surgical treatment of acute purulent mediastinitis A. Cherveniakov 1 and P. Cherveniakov’ 1 Emergency Medical Institute “N. I. Pirogov”, ’ Bulgarian Medical Academy, Sofia, Bulgaria

Sofia, Bulgaria

Abstract. During the last 15 years we have treated 147 patients with acute purulent media&i&is (APM). According to the aetiology of the disease, 2 major groups were defined. The first group included the cases of oesophageal origin - 112 patients (dilatation - 38 patients, foreign body extraction - 29 patients, lye injuries - 11 patients, oesophagoscopy - 8 patients, sharp foreign body - 6 patients). The second group consisted of patients with mediastinitis of non-oesophageal origin - 35 patients (trachea-bronchial disease - 21 patients, tooth infection - 8 patients, cervical infection). Symptoms of the mediastinal infection were typical; nevertheless, early diagnosis (within first 12 h) was obtained in only 43.5% of cases. Therapy for all patients included general stabilisation, broad spectrum antibiotics and immunotherapy. In 86 patients, mediastinal drainage was performed with additional suture of the oesophageal wall or plication with a gastric or diaphragmatic patch in 9 cases. Oesophagectomy and delayed colon transplant was the method used in 61 patients. Mortality included 21 patients (14.3%). The cause was broncho-pneumonia in 9 patients, endotoxic shock in 7 and renal failure in 3 patients. [Eur J Cardio-thorac Surg (1992) 6:407-4111 Key words: Mediastinitis ~ Descending cervical abscess - Oesophageal perforation

The incidence of acute purulent mediastinitis has been increasing during the last two decades according to the medical literature. The widespread use of oesophageal endoscopic investigations and treatment has raised the number. The prognosis is variable, depending on the condition of the oesophageal wall, the type of injuries and the interval before diagnosis. The number of reported cases with mediastinal infection from diseases of the tracheobronchial tree, neck and thoracic operations also increased. The main principles of treatment were described by Hippocrates - “Any pus must be evacuated”. Modern antibiotics, parenteral nutrition, resuscitation as well as surgical techniques have given a new therapeutic contribution to the treatment of the disease. An individual approach in the treatment of every patient is preferable. Many authors give priority to either predominantly conservative treatment or to an active surgical intervention. Unfortunately, a high mortality and morbidity remains and mediastinitis is still a life-threatening condition. By sharing our experience we are looking forward to further exchanges of ideas in the search for optimal criteria and principles of treatment. Presented at the 5th Annual Meeting of the European Association for Cardio-thoracic Surgery, London, UK, September 23-25, 1991

Material and methods At the Clinic for Thoracic Surgery in the Bulgarian Medical Academy and the Emergency Medical Institute “N.I. Pirogov”, Sofia, we have treated 147 patients with acute purulent mediastinitis (APM) over a period of 15 years. Their mean age was 37 years and 75.5% were male (Table 1). According to the cause of the inflammation, our patients were divided into two main groups. In 112 patients, mediastinitis was due to oesophageal perforation. Dilatation of a stricture (n=38), extraction of a foreign body (n=29) and lye injuries (n = 11) were the most common causes of perforation (Table 2). A diseased oesophageal wall was confirmed in 64,3% (n =72) of the patients; in 35.7% (n=40) it was intact. The majority of oesophageal perforations were intrathoracic but in 19 cases (17% of the first group) they were cervical. Pleural empyema occurred in 28 patients, 6 being bilateral. In 9 patients with a perforation above the cardia, mediastinitis was combined with peritonitis. In the second group of 35 patients, mediastinitis was due to non-oesophageal disease. The infection in this group came from

Table 1. Clinical details

Sex

Number

Mean age”

Male Female

111 (75.5%) 36 (24.5%)

37 38

a The age ranged from 1.5 to 82 years

408 Table 2. Acute purulent

mediastinitis

of oesophageal

Dilatation of stricture Extraction of foreign body Severe lye injuries Oesophagoscopy Sharp foreign body Chest injury Leak of neck anastomosis Interposition of endoprosthesis Necrosis of cancer Necrosis of transplant: Colon Jejunum Perforation of transplant peptic ulcer

origin 38 29 11 8 6 5 4 4 3 2 1 1

Total

112 patients

Table 3. Mediastinitis

of non-esophageal

origin

Fig. 1. Chest X-ray of patient with mediastinitis caused by descending cervical abscess form tracheostomy following respiratory arrest. Bilateral cervical incision with cleansing of the mediastinum was performed followed by mediastinal irrigation for a week

Trachea or broncho-oesophageal tistula Odontogenic infection After lung resection Post intubation lesions of the trachea Cervical abscess Thyroid gland abscess

12 8 6 3 3 3

Total

35 patients

Table 4. Time of the diagnosis of the mediastinitis Early (first 12 h) Late (between 12 h-72 h) Late (after more than 72 h)

Table 5. Treatment

64 patients (43.5%) 83 patients (50.25%) 7 patients (6.25%)

of acute purulent mediastinitis

1 Medical treatment a Broad spectrum antibiotics b Intravenous hydratation, fresh blood, plasma, parenteral feeding c Immunotherapy: hyperimmune and antistaphilococcal plasma, gamma globulin, gamma venin d Respiratory physiotherapy and resuscitation 2 Surgical technique a Drainage and lavage of the mediastinum b Primary closure of the oesophageal wall c Oesophagectomy and gastrostomy with substitution of the oesophagus within 3 months

Table 6. Postoperative

morbidity

primary maligant or benign disease of the tracheobranchial tree, most often involving the oesophagus (n = 15), from tooth infections (n = 8), after lung resection (n= 6), or from cervical (n= 3) and thyroid (n = 3) infections (Table 3). The development of mediastinitis was characterized in most patients by the classic symptoms of chest pain, pyrexia with toxic features, oedema, emphysema, tachycardia and dyspnoea with associated leucocytosis. Only 14 patients presented with atypical features of the disease. Plain chest radiography usually confirmed the diagnosis. Distinct evidence included an enlarged mediastinum, mediastinal emphysema, air fluid levels, pleural effusion (unilateral or bilateral) and foreign bodies. They were observed in two thirds of the patients. A sterile contrast swallow was performed to prove oesophageal perforation in all patients with mediastinal and pleural space leak. Endoscopic examination was performed when previous data were not sufficient. This was the method of choice in cases with fistulae of the trachea-bronchial tree and in cases with early perforation. Unfortunately, early diagnosis within the first 12 h in nonspecialized units, was obtained in only 43.5% of cases (n= 64) and in 7 patients there was a diagnostic delay greater than 72 h (Table 4). A mixed aerobic flora was the commonest culture. Mixed aerobic and anaerobic agents were found in 45 patients (30,6%).

y:g; n

Results

(n=61)

Medical treatment was started immediately when acute mediastinal infection was suspected. At first, a massive dose of broad spectrum antibiotics was given and revised according to the microbiology results. Intravenous hydratation, and in oesophageal patients, parenteral feeding was used. In delayed and septic patients, immunotherapy with hyperimmune plasma, gamma globulin and gamma venin was added. Respiratory physiotherapy and early enteral feeding through naso-gastric or gastrostomy tubes were added. An individual approach in the selection of surgical procedure was performed. Mediastinal drainage and irrigation were generally used in patients with nonoesophageal APM or in the first group of patients, in the early stage of perforation, with an intact oesophageal

and mortality

TN=

Number of patients

Cured

Deaths

Bronchopneumonia Septic shock Acute renal failure Deep vein thrombosis Lung abscess

24 8 4 2 2

14 1 1 I 1

9 7 3 1 1

Total

40

18

21 (14.4%)

409

wall and with a predominantly cervical localisation (n = 89, (Fig. 1). Primary closure of the tear was the method of choice in 29 patients in good general condition without significant local alterations. Suture of the intrathoracic oesophagus was reinforced with a parietal pleural flap. In subdiaphragmatic tears, this was reinforced with an omental graft. In 3 cases with supracardial perforation after dilatation of a benign stricture, we successfully used a plastic reconstruction with a wedge stomach flap (Fig. 2). In patients with a late diagnosis of oesophageal perforation, advanced symptoms of mediastinitis, pleural empyema or pneumoperitoneum, a major surgical procedure was performed. The first stage was resection of the oesophagus with a cervical oesophagostomy and gastrostomy plus mediastinal toilet (n = 72). Retrosternal colon substitution of the oesophagus was performed within 3 months after primary operation in 61 patients (Table 5). Early survival was 85,6%. One hundred and twentysix of our patients were discharged following recovery. The morbidity in the early postoperative period (30 days) was 27% (40 patients). Twenty-one patients (14.4%) died. Factors affecting mortality included severe lye injuries - 7/11 cases (63% mortality), inoperable carcinoma of the oesophagus (2/3 cases), and previous oesophageal substitution (2/4 patients). In all complicated cases treatment had been started later than 12 h after the onset of symptoms. Bronchopneumonia was the most common complication, followed by septic shock and renal failure (Table 6). Extension of the infection to the peritoneal cavity increased the mortality (about 50% in cases with peritonitis and bilateral pleural empyema). The patients aged over 65 in poor general condition and with accompanying diseases were also prone to complications. Extensive surgical treatment did not alter the outcome. Of 72 patients with resections of the oesophagus, 61 survived and underwent substitution. Discussion

Over the last decade, the incidence of APM has increased. Most authors, with a single exception [15], noted the

same phenomenon [2, 11, 12, 13, 141. In recent reports, mediastinitis does not seem to be a common disease, but it still remains a life-threatening condition [5, 6, 7, 9, 10, 191with a mortality varying from 100% to 0%. The overall mortality still remains around 30% [2, 12, 16, 18, 191. The important factors of early diagnosis and successful treatment are interrelated in direct proportion. This is the reason why patients with APM should be divided into two groups: a) begin of treatment within 24 h and b) after that time [2, 151. Delay in diagnosis was due firstly to failure to recognise the symptoms of the disease in 55% of our cases (n = 81), secondly to misinterpretation of the chest X-rays (enlarged mediastinum, mediastinal emphysema in 25% (n = 37) and thirdly to atypical mild symptoms of the disease in 9.5% (n = 14). The clinical symptoms of APM are very important and they can be enough to confirm the diagnosis for exploration even in false negative studies [2]. In our experience, mediastinal infection after 12 h without treatment is already advanced with local infiltration, mixed bacterial flora and a tendency to expansion. Development of pneumothorax, pneumoperitoneum, mediastinal emphysema, sepsis and shock are most often symptoms of delayed treatment [I 11. The value of early medical treatment was proved in practice and is performed routinely, but the surgical treatment of APM still remains controversial. The surgical procedures must be individualised according to the findings of the specitic case [2, 6, 12, 20, 211. We do not consider that expectanteonservative treatment can succeed in cases with pus in the mediastinum. The correct place for that treatment is in cases with localised intramural perforation of the oesophagus [I, 91. The aim of surgical management is cleansing of the mediastinum and a definitive surgical procedure at the source of the sepsis. Performance of extensive drainage and lavage is successful in patients with localised tooth infection, cervical abscess and cervical perforation. Intrathoracic or subdiaphragmatic perforation of the oesophagus or infection from other origins are indications for surgical intervention. The oesophagus has a tendency to heal [4, lo] but in our cases, operation was necessary for mediastinal inflammation. Local drainage and primary closure are pos-

410

sible in the early stage of oesophageal perforation with localised inflammation [8, 193. Routine performance of pleural flaps is useful. We appreciate the advice of J. R. Benfield to the effect that “The risk of doing too little, too late, is death” [3]. Resection of the oesophagus, cervical oesophagostomy and gastrostomy with second stage substitution is the method of choice in patients with advanced APM of oesophageal origin with a tendency to expansion and generalisation of the infection. The overall mortality was not influenced by the type of surgical method chosen. Excellent results with primary closure of the tear in the early stage after iatrogenic perforation of the oesophagus determined our approach in patients with a fixed sharp foreign body in the gullet. In 9 patients, thoracotomy, oesophagectomy and extraction of the foreign body was performed without complications. In conclusion, APM is a disease which can be easily missed in an early phase but also can be more successfully treated at this stage.

References Berliner L, Redmond P, Patchter L (1982) Spontaneous intramural perforation of the esophagus. Case report and review of the literature. Am J Gastroenterol 77:355-357 Bladergroen MR, Lowe JE, Postlethwait RW (1986) Diagnosis and recommended management of esophageal perforation and rupture. Ann Thorac Surg 42~235-239 Brewer LA, Carter R, Mulder GA, Stiles R (1986) Options in the management of perforations of the esophagus. Am J Surg 152:62-69 Cameron JL, Kieffer RF, Hendrix TR, Mehigan DG, Barker R (1979) Selective nonoperative management of contained intrathoracic esophageal disruptions. Ann Thorac Surg 27:404-408 DeMeester Tom R (1986) Perforations of the esophagus. Ann Thorac Surg 42:231-232 6. Flynn AE, Verrier ED, Way LW, Thomas AN, Pellegrini CA, (1989) Esophageal perforation. Arch Surg 124: 1211-1215 d.

7. Fukamoto K, Sugita R (1990) A case of cervico-mediastinal abscess secondary to acute tonsillitis: investigation and treatment. Nippon Jibiinkoka Gakkai Kaiho 93:884-893 8. Gouge TH, Depan HJ, Spenser F (1989) Experience with Grillo pleural wrap. Ann Surg 209:612-619 9. Jamieson GG (1989) Spontaneous and iatrogenic perforation of the oesophagus. Surgery 1796: 1798 10. Lyman A, Brewer, Carter R, Mudler GA, Stiles QR (1986) Options in management of perforations of the esophagus. Am J Surg 152:62-69 11. Lyons WL, Seremetis MG, decuzman VC, Peabody JW (1978) Ruptures and perforations of the esophagus: the case for conservative supportive management. Ann Thorac Surg 25:346-350 12. Luc Michel, Grill0 HC, Malt RA (1982) Esophageal perforation. Ann Thorac Surg 32203-210 13. Merkle NM, Foitzik Th, Riedl S, Fritz Th (1990) Infektionen des Mediastinums. Chirurg 61:629-638 14. Moghissi K, Pender D (1988) Instrumental perforations of the oesophagus and their management. Thorax 43: 641-646 15. Petterson G, Larson S, Gatzinsky P, Sudow G (1981) Differentiated treatment of intrathoracic oesophageal perforations. 15:321-324 16. Sawyers JL, Lane CF, Foster JH, Daniel RA (1975) Esophageal perforation. Ann Thorac Surg 3:233-238 17. Seaman M, BaIIinger P, Sturgill TD, Maertins M (1991) Mediastinitis following nasal intubation in the emergency department. Am J Emerg Med 9:37-39 18. Scinner DB, Little AG, DeMeester TR (1980) Management of esophageal perforation. Am J Surg 139:760-764 19. Slater G, Scular AA (1982) Esophageal perforations after forceful dilatations in achalasia. Ann Surg 195: 186-188 20. Tilanus HW, Bossuyt P, Schattenkerk ME, Obertop H (1991) Treatment of oesophageal perforations: a multivariate analysis. Br J Surg 78: 582 - 585 21. Wheatley MJ, Stirling MC, Kirsh MM, Gago 0, Orringer MB, (1990) Descending necrotizing mediastinitis: transcervical drainage is not enough. Ann Thorac Surg 49:780-784

Prof. Peter Cherveniakov Bulgarian Medical Academy bul. F. Nansen 45 Sofia 1042 Bulgaria

411

Discussion Dr. Pastore (Naples, Italy). Thank you for the opportunity of letting me comment on this paper of considerable interest. I completely agree with the conclusion of Dr. Cherveniukov who indicates the need for early diagnosis and aggressive management of purulent mediastinitis. Early diagnosis, antibiotic therapy and adequate surgical debridement of the mediastinum is the key to a successful outcome. The diagnosis of mediastinitis can be difficult, the plain roentgenograms are often non-diagnostic until sepsis has developed. The CT scan provides an earlier and accurate test for diagnosing the presence of mediastinitis and is extremely useful in following the adequacy of surgical drainage. In the patients in whom esophageal perforation cannot be repaired, we prefer esophageal diversion, proximal exclusion or stenting across the perforation to eliminate the mediastinal contamination. In my opinion, the indication of esophagectomy must be limited and primarily directed toward extensive esophageal devitalization. I have two questions. Do you consider using these conservative procedures as alternatives to esophagectomy. Second, in more extensive descending mediastinitis of nonesophageal origin where the transcervical mediastinal drainage is often not enough, what other surgical approaches would you use for complete management of infection? I would also like to thank you for sending me the manuscript. Dr. M. Ribet (Lille, France). Please accept my congratulations because I think in such a series it is very rare to obtain this rate of success. A 14% mortality is a result to be commended by this Association. Mr. B. Ross (Norwich, Mx). I am very interested in this paper. I was a bit surprised that in the whole paper you didn’t mention spontaneous rupture. All your esophageal problems seem to be related to iatrogenic causes. Of course, prevention is better than cure. I saw those beautiful slides you showed with barium trickling in the mediastinum obviously related to dilatation procedures which I presume was with rigid instruments. I would suggest that if you use X-ray control guide wires and perhaps Maloney, or in a better position dilators, the type of perforation that you show would not occur. It was also interesting to see that you mentioned gastrostomy. I presume that you were using a feeding gastrostomy. I think this actually perpetuates reflux and contaminates the perforation. The gastrostomy should be a draining procedure and you should feed through a jejunostomy. Dr. G. Pettersson (Copenhagen, Denmark). I want to congratulate you to your results with these very difficult patients. There is, how-

ever, a number of patients with perforation of the esophagus and mediastinitis, who have no underlying strictming esophageal disease. In nine advanced such cases, we have drained the esophagus by a T-tube after surgical debridement of the mediastinum and pleura. All patients had an uncomplicated postoperative course and recovered normal esophageal function. [Larsson S. Pettersson G, Lepore V (1991). Esophagocutaneous drainage to treat late and complicated esophageal perforation. Eur J Cardio-thorac Surg 557995821. Finally I would like to ask you in how many patients without any underlying esophageal disease did you perform esophagectomy? Prof. J. Vogt-Moykopf (Heidelberg, FRG). You mentioned the importance of perioperative treatment with antibiotics which is certainly essential. In patients without any immunodeficiency we do not give globulin, gamma globulin and so on routinely. Could you explain your reasons why you do so, because this is very expensive medicine. Mr. K. M. Pagliero (Exeter, Mx). I would like to congratulate the authors of this paper and commend what they suggest. However, every patient needs to be treated on his merits. I should like to mention an alternative intervention which we have had published [Ann Thorac Surg (1991) 52: 151- 1531 in the past couple of months whereby we have shown it may be of benefit in selected cases, rather than to drain externally via a T-tube, to drain internally via a Salem sump drain that is introduced via a pharyngostomy, into the oesophagus, through the perforation and into the mediastinum. This is particularly suitable if the perforation is limited to the mediastinum without breach of the pleura. Dr. Cherveniakov: Thank you all for discussing our paper. In answer to Dr. Pastore’s question, oesophagectomy was performed for the following reasons. At first in our experience there was quite a large number of perforations in early or late stages following lye injuries of the oesophagus. Over 30 of them underwent oesophagectomy. The other significant group of patients undergoing oesophagectomy was those with delayed diagnosis. Because of the high risk of oesophageal repair, radical surgery was performed. In answering to the next question, we performed 72 oesophagectomies and in 61, colon substitution was used at a later date. Concerning the cervical drainage of the mediastinum, we have also had several cases where this was insufficient. Our method of choice was not median stemotomy because of the risk of severe sternal inflammation. We performed left or right thoracotomy to treat the mediastinitis. In cases of enteral feeding, a jejunostomy tube can be used. In our experience the new feeding solutions were not available but gastrostomy feeding proved adequate. Concerning the two questions about the T-tube, we have no experience with this method, but as we mentioned the treatment of the patients must be strictly individualized and inserting of the T-tube can be the method of choice in some cases.