Eur J Cardio-thorac
Surg (1992) 6561-564
c, Springer-Verlag
1992
Antibiotic prophylaxis in oesophageal surgery D. A. C. Sharpe, P. Renwick, K. H. R. Mathews, and K. Moghissi North Humberside
Cardiothoracic
Centre, Castle Hill Hospital, Cottingham,
North Humberside.
UK
Abstract. A prospective random&d study to assess the efficacy of antibiotic prophylaxis in oesophageal surgery was performed, in which 226 consecutive patients (113 male and 113 female, age range 24-86 years, mean age 65 years) were included. The study patients were in two groups: group 1, in which the upper alimentary tract was opened during surgery, and group 2, in which it was not. The group 1 patients (II = 129) were randomised to one of three antibiotic prophylaxis regimes prior to surgery. Group A patients (n = 42) were treated with cefuroxime (CFX) 1.5 g at induction of anaesthesia and then CFX 750 mg b.i.d. for 4 days. Group B patients (n = 46) were treated with CFX 1.5 g and metronidazole (MTR) 1.0 g at induction of anaesthesia, then CFX 750 mg b.i.d. and MTR 500 mg qds for 4 days. Group C (n = 41) treated with CFX 1.5 g and MTR 1.0 g at the induction of anaesthesia. Group 2 (n = 97) was divided into two groups, group D (n = 47) treated with CFX 1.5 g on induction of anaesthesia alone. Group E (II = 50) treated with CFX 1.5 g on induction of anaesthesia then CFX 750 mg bd for 2 days. We found a significantly higher incidence of infective complications in subgroup C (43.9%) and subgroup A (21.4%) compared to subgroup B (8.6%). This difference was most marked in patients undergoing oesophagectomy. We found significantly higher infection rates of infective complications in subgroup D (10.6%) as compared to subgroup E (2%). [Eur J Cardio-thorac Surg (1992) 6:561-5641 Key words: Oesophagus - Surgery - Infection - Antibiotics
Surgical resection offers the best hope for long-term survival and provides effective palliation in patients with carcinoma of the oesophagus. However, surgical procedures for the resection of oesophageal carcinoma have an in-hospital mortality in the order of 10% [13]. A significant cause of this considerable mortality is generally agreed to be the infective complications of surgery. Inhospital mortality has improved over the last 10 years, and one of the factors considered to be responsible for this improvement is antibiotic prophylaxis [3]. Numerous studies demonstrate the efficacy of prophylactic antibiotics in the surgery of the lower gastrointestinal tract [2,4,15], but relatively few studies have been undertaken to assess the efficacy of antibiotic prophylaxis in surgery of the upper alimentary tract. This study was set up prospectively to test the efficacy of antibiotic prophylaxis regimes in patients undergoing major surgery to the upper gastrointestinal tract.
Received for publication: Accepted for publication:
November 7, 1991 April 6, 1992
Patients and methods The study was organised as a prospective randomised trial. There were no specific criteria for exclusion from the study, and all patients undergoing major oesophageal surgery during the 2.5 years of the study were included. All patients were under the care of a single consultant surgeon, all operations being performed by his team. Two hundred and twenty-six patients were entered into the study, 113 male and 113 female. The age range was from 24 to 86 years, mean 65.3 years. The study patients were divided into two distinct groups, those in whom the gastrointestinal tract was opened as part of the surgical procedure (group l), and those in whom it was not (group 2). In group 1 (n = 129) the patients had a mean age of 67 years (range 24-86 years) and consisted of 51 female and 78 male patients. This group contained all of the patients suffering from oesophageal carcinoma (n = 105), and these patients were distributed evenly throughout the antibiotic regime subgroups A, B, and C (see below). For the operations performed, see Table 3. Oesophagectomy was performed by one of three techniques: laparotomy and right thoracotomy, a left thoracophrenotomy approach, or a transhiatal technique [I, 31, 141. All patients undergoing oesophagectomy except a single patient in subgroup A were suffering from oesophageal cancer. In group 2 (n = 97) the patients had a mean age of 61.8 years (range 25-84 years) and consisted of 62 female and 35 male patients. The operations performed were either an antireflux proce-
562 dure, with a Moghissi modified total fundoplication [12] or myotomy for achalasia of the cardia [5]. All study patients in group 1 were randomised at the time of operation to one of three antibiotic regimes. Group A (n = 42) were treated with cefuroxine (CFX) 1.5 g at induction of anaesthesia and then CFX 750 mg b.i.d. for 4 days. Group B (n = 46) were treated with CFX 1.5 g and metronidazole (MTR) 1.0 g at induction of anaesthesia, then CFX 750 mg b.i.d. and MTR 500 mg qds for 4 days. Group C (n = 41) were treated with CFX 1.5 g and MTR 1.O g at the induction of anaesthesia and no further antibiotics unless clinically indicated. Group 2 patients were randomised to one of two antibiotic regimes: group D (n = 47) were treated with CFX 1.5 g on induction of anaesthesia alone, while group E (n = 50) were treated with CFX 1.5 g on induction of anaesthesia, then CFX 750 mg b.i.d. for 2 years. The criteria for infection used in the study were as follows: - Wound sepsis: a discharging surgical wound containing pus, with or without pathogens. _ Pulmonary infection: the presence of clinical (fever, leucocytosis) and radiological signs of pulmonary infection with sputum containing pus, with or without pathogens. - Pleural infection: the presence of pus in the chest (empyema). _ General infection: clinical signs of infection in association with blood cultures yielding pathogens. All patients had routine samples of sputum sent for microbiological examination pre-operatively and on the 3rd, 7th and 11th post-operative days. Wounds were inspected daily for evidence of infection. Any potential episode of infection was investigated using an infection screen consisting of sputum culture, wound swab, urine culture and blood culture. If any other microbiological investigations were considered necessary they were ordered on an individual basis. When an infective complication occurred, administration of CFX was either increased to the treatment dose (CFX 750 mg tds) or restarted at this dosage. MTR was given if still in use as part of the prophylaxis regime, but if it had already been discontinued it was not restarted routinely for infective complications. Treatment with CFX was continued until definitive microbiological results from the infection screen were available, then the antibiotic regime altered accordingly if necessary. Any infective complication occurring during the hospital admission for surgery was considered in this study. All patients were routinely screened with contrast studies on the 4th (group 2) or the 7th post-operative day. Results were analysed using x ’ tests and contingency tables.
Table 1. Organisms identified from the pre-operative sputum samples and the number of cases in which they were isolated” Organisms
isolated
n
Normal flora Haemolytic streptococcus Hemophilus injluenzae Streptococcus pneumoniae Citrobacter sp. Enterobacter sp. Escherichia coli Klebsiella sp. Proteus sp. Pseudomonas sp. Staphylococcus aureus Streptococcus faecalis Candida albicans
192
14 13
a Thirteen patients had more than one organism their pre-operative sputum samples
Table 2. Distribution Antibiotic regime (induction/ post-op.) Group 1 A CFX/CFX B
CFX/CFX MTR/MTR
C
CFX MTR
Group 2 D CFX E
CFX/CFX
of the infective complications Infections ~ n %
identified
from
by group
Site Lungs
Wound
9
21.4
4
3
4
8.6
2
1
18
43.9
13
4
5
10.6
2
3
2
2
Genera1
Pleura
2 1 1
2
CFX, Cefuroxime; MTR, Metronidazole ’ Three patients in group C suffered infection in more than one site
Results
Within the study group of patients there were no instances of oesophageal leakage and no in-hospital death. The pre-operative sputum samples showed potentially pathogenic organisms in 15% of cases. The distribution and frequency of each the isolates from the preoperative sputum samples is shown in Table 1. We could find no correlation between the pathogenic organisms identified from the pre-operative sputum samples and the organisms identified from any subsequent infective complications (P > 0.05). The surgical technique employed for each type of procedure in each of the two study groups for any of the antibiotic prophylaxis regimes made no significant difference to the infective complication rate (P > 0.05). Group 1 patients suffered a total of 31 infective episodes during the study, while group 2 patients suffered 7 infective episodes: the difference was significant (P < 0.05). The most frequently identified infection was pulmonary sepsis: 19 cases in group 1 and 2 cases in group 2. The higher incidence of pulmonary sepsis in
group 1 is a highly significant result (P < 0.001). In association with pulmonary sepsis, three patients in group 1 also developed empyema; in each of these cases the same organism was cultured from both sites. There were 13 wound infections, 8 in group 1 and 5 in group 2; the higher incidence within group 1 is not significant (P > 0.05). A single patient in group 1 suffered from generalised sepsis. In 92% of cases the isolated organism was sensitive to one of the antibiotics employed for prophylaxis. Table 2 shows the incidence of infective complications with respect to the antibiotic subgroups. In group 1 the incidence of infection was significantly higher in groups A (P < 0.05) and C (P c 0.001) than in group B. When pulmonary sepsis and wound infections were considered we found the same significantly higher incidence in groups A (P < 0.05) and C (P < 0.001) than in group B. Significantly more infections occurred in group C than in groups A (P c 0.05) and B (P < 0.05). In Table 3 incidence of infection is considered by type of operation involved as well as by antibiotic subgroup. Patients who underwent oesophagectomy had a sign&
563 Table 3. Distribution of surgery involved
of infective complications
by group and type
Number of patients
Number with infections
26 35 31
6 (23%) 3 (8.5%) 14 (45%)
Group and operation
Group 1 Oesophagectomy A B C Intubation A B C
(n = 14)
Oesophagoplasty A B C Group 2 Antireflux D E Myotomy D E
(n = 92)
8
1
3 3
2
8 8 7
2 1 2
37 41
3 2
10
1 1
(n = 23)
(n = 78)
(n = 19) 9
cantly higher incidence of infective complications than those undergoing intubation (P < 0.001) or oesophagoplasty (P < 0.05). Within the oesophagectomy subgroup there was a significantly higher incidence of infective complications in groups A (P < 0.05) and C (P < 0.001) as compared to group B. In regard to the other procedures performed, the groups are too small for meaningful statistical analysis. It should be noted that in both the oesophagoplasty and the oesophageal intubation groups, infective complications were more frequent in subgroup C. In group 2 we found a significantly higher incidence of infection in subgroup D (P < 0.05) than in subgroup E. In this group the complication rate for the various types of infection was too low for meaningful statistical analysis to be performed. The operative technique involved did not affect the infective complication rate (P > 0.05). Discussion
The use of antibiotic prophylaxis is now considered to be routine in most major surgical procedures on the alimentary tract, and there are numerous studies which support this viewpoint [2, 4, 151, although it should be stressed that antibiotic prophylaxis is never a substitute for good surgical practice and meticulous aseptic technique. So far, no other clinical trials have considered the question of antibiotic prophylaxis for major oesophageal surgery. The study groups reflect the epidemiology of the principal disease types in each of the groups: group 1 contained all patients with carcinoma of the oesophagus and thus the greater number of male patients in the study, while group 2 consisted principally of patients with reflux disease and therefore had the greater number of female patients.
CFX was chosen as our antibiotic for prophylaxis in this study because it has a broad spectrum which would provide adequate prophylaxis against the potential pathogenic organisms, and because of its good penetration into the likely sites of sepsis [9]. In subgroups B and C, MTR was added to enhance the anaerobic cover provided, in view of the spectrum of organisms likely to be encountered [IO]. These suppositions were confirmed by the spectrum of organisms encountered, most of which were sensitive to be antibiotic regime employed. In group 1 we found a considerably higher infective complication rate than in group 2. The incidence of pulmonary sepsis in group 1 was similar to that found in the studies of Fan, Keeling and Galanduik et al. [6,7,8]. The higher infective complication rate within group 1 reflects a number of factors. First, in this group the alimentary tract was opened, and therefore the operative field must be considered to be contaminated. Secondly, the general condition of the patients in group I, a high proportion of whom were suffering from malignant disease, was not as good as in group 2. Lastly, in general patients underwent more extensive surgical procedures, and our results show that the patients undergoing oesophagectomy sustained more infective complications than those undergoing other operative procedures. The conclusion from this is that resectional surgery of the oesophagus is the greatest single risk factor for infective complications, and that the more extensive the operative procedure, the greater the risk of an infective complication. Within group 1 we found a lower incidence of infective complications in subgroup B than in the other two subgroups. Since the subgroups were comparable, the only difference being the antibiotic prophylaxis regime, it must be concluded that the antibiotic prophylaxis given to subgroup B was the most efficacious. This conclusion is true in regard to pulmonary sepsis and wound infection. If the patients undergoing oesophagectomy are considered separately the results are much more marked, with subgroup B clearly having superior prophylaxis to groups A and C. The higher incidence of infective complications within the oesophagectomy subgroup appears to accentuate the superiority of this antibiotic prophylaxis regime. For oesophagoplasty and intubation we did not have sufficient numbers to statistically prove the superiority of the antibiotic prophylaxis regime in subgroup B, although the incidence of infective complications was still lowest in this subgroup. Within group 2 there was a significantly lower incidence of infective complications in subgroup E than in subgroup D. The numbers involved are too low to perform meaningful statistical analysis in regard to the various sites of infection. The operative procedure employed appeared to have no role to play in the frequency of infective complications in group 2.
Conclusions
1. Adequate antibiotic prophylaxis for surgery to the upper gastrointestinal tract when the lumen of a viscus is to be opened is provided by CFX 1.5 g and MTR 1.0 g at
564
induction of anaesthesia, then CFX 750 mg b.i.d. and MTR 500 mg qds for 4 days. 2. Adequate antibiotic prophylaxis for surgery to the upper gastrointestinal tract when the lumen of a viscus is not opened is provided by CFX 1.5 g on induction of anaesthesia, then CFX 750 mg b.i.d. for 2 days. As a resuh of this study, the antibiotic prophylaxis regimes outlined above are now those used in our unit.
7.
8. 9.
IO. Acknowledgement. We would like extend thanks to Dr Meigh and the staff of the microbiology department of Castle Hill Hospital for their assistance in this study.
11.
References
12.
1. Churchill ED, Sweet RH (1942) Transthoracic resection of tumours of the stomach and oesophagus. Ann Surg 115: 897 -917 2. Condon RE, Bartlett JG, Greenlee H, Schulte WJ, Ochi S, Abbe R, Caruana JA, Gordon HE, Horsley JS, Irvan G 3rd, Johnson W, Jordan P Jr, Keitzer WS, Lempeke R, Read RC, Schumer W, Schwartz M, Storm FK, Vetto RM (1983) Efficacy of oral and systemic antibiotics prophylaxis in colorectal surgery. Arch Surg 118: 496 3. Earlam RJ, Cunha-Melo JR (1980) Oesophageal squamous cell carcinoma. 1. A critical review of surgery. Br J Surg 67: 381-390 4. Edmondson HT, Rissing JP (1983) Prophylactic antibiotics in colon surgery. Arch Surg 118: 217 5. Ellis FH Jr, Olsen AM, Schlegel JF, Code CF (1964) Surgical treatment of oesophageal hypermotility disturbances. JAMA 188: 862-866 6. Fan ST, Lau WY, Yip WC, Poon GP, Yeung C, Lam WK, Wong KK (1987) Prediction of post operative pulmonary com-
13.
14. 15.
plications in oesophogastric cancer surgery. Br J Surg 74: 408410 Galanduik S, Hermann RE, Gassman JJ, Cosgrove DM (1986) Cancer of the oesophagus: the Cleveland clinic experience. Ann Surg 203: 101-108 Keeling P, Gillen P, Hennessy TP (1988) Oesophageal resection in the elderly. Ann R Co11 Surg Engl 70: 34-36 Kucers A, Bennet N (1979) Cefamandole, cefoxitin, cefuroxime and ceforanide. In: Kucers A, Bennet N (eds) The use of antibiotics. Heinemann, London, pp 267-293 Kucers A, Bennet N (1979) Metronidazole. In: Kucers A, Bennet N (eds) The use of antibiotics. Heinemann, London, pp 761-781 Lewis I (1946) The surgical treatment of carcinoma of the oesophagus with special reference to growths of the middle third. Br J Surg 34: 18 Moghissi K (1986) Essentials of thoracic and cardiac surgery. Heinemann, London, pp 232-233 Muller JM, Erasmi M, Stelzner M, Zieren U, Pichlmaier H (1990) Surgical therapy of oesophageal carcinoma. Br J Surg 77: 845- 857 Orringer MB, Sloan H (1978) Esophagectomy without thoracotomy. J Thorac Cardiovasc Surg 76: 643 -654 Portnoy J, Kagan E, Gordon PH, Mendelson J (1983) Prophylactic antibiotics in elective colorectal surgery. Dis Colon Rectum 26: 310
Mr D. A. C. Sharpe, MB, FRCS North Humberside Cardiothoracic Centre Castle Hill Hospital Castle Road Cottingham, North Humberside HU16 5JQ United Kingdom
Surg (1992) 6561-564
c, Springer-Verlag
1992
Antibiotic prophylaxis in oesophageal surgery D. A. C. Sharpe, P. Renwick, K. H. R. Mathews, and K. Moghissi North Humberside
Cardiothoracic
Centre, Castle Hill Hospital, Cottingham,
North Humberside.
UK
Abstract. A prospective random&d study to assess the efficacy of antibiotic prophylaxis in oesophageal surgery was performed, in which 226 consecutive patients (113 male and 113 female, age range 24-86 years, mean age 65 years) were included. The study patients were in two groups: group 1, in which the upper alimentary tract was opened during surgery, and group 2, in which it was not. The group 1 patients (II = 129) were randomised to one of three antibiotic prophylaxis regimes prior to surgery. Group A patients (n = 42) were treated with cefuroxime (CFX) 1.5 g at induction of anaesthesia and then CFX 750 mg b.i.d. for 4 days. Group B patients (n = 46) were treated with CFX 1.5 g and metronidazole (MTR) 1.0 g at induction of anaesthesia, then CFX 750 mg b.i.d. and MTR 500 mg qds for 4 days. Group C (n = 41) treated with CFX 1.5 g and MTR 1.0 g at the induction of anaesthesia. Group 2 (n = 97) was divided into two groups, group D (n = 47) treated with CFX 1.5 g on induction of anaesthesia alone. Group E (II = 50) treated with CFX 1.5 g on induction of anaesthesia then CFX 750 mg bd for 2 days. We found a significantly higher incidence of infective complications in subgroup C (43.9%) and subgroup A (21.4%) compared to subgroup B (8.6%). This difference was most marked in patients undergoing oesophagectomy. We found significantly higher infection rates of infective complications in subgroup D (10.6%) as compared to subgroup E (2%). [Eur J Cardio-thorac Surg (1992) 6:561-5641 Key words: Oesophagus - Surgery - Infection - Antibiotics
Surgical resection offers the best hope for long-term survival and provides effective palliation in patients with carcinoma of the oesophagus. However, surgical procedures for the resection of oesophageal carcinoma have an in-hospital mortality in the order of 10% [13]. A significant cause of this considerable mortality is generally agreed to be the infective complications of surgery. Inhospital mortality has improved over the last 10 years, and one of the factors considered to be responsible for this improvement is antibiotic prophylaxis [3]. Numerous studies demonstrate the efficacy of prophylactic antibiotics in the surgery of the lower gastrointestinal tract [2,4,15], but relatively few studies have been undertaken to assess the efficacy of antibiotic prophylaxis in surgery of the upper alimentary tract. This study was set up prospectively to test the efficacy of antibiotic prophylaxis regimes in patients undergoing major surgery to the upper gastrointestinal tract.
Received for publication: Accepted for publication:
November 7, 1991 April 6, 1992
Patients and methods The study was organised as a prospective randomised trial. There were no specific criteria for exclusion from the study, and all patients undergoing major oesophageal surgery during the 2.5 years of the study were included. All patients were under the care of a single consultant surgeon, all operations being performed by his team. Two hundred and twenty-six patients were entered into the study, 113 male and 113 female. The age range was from 24 to 86 years, mean 65.3 years. The study patients were divided into two distinct groups, those in whom the gastrointestinal tract was opened as part of the surgical procedure (group l), and those in whom it was not (group 2). In group 1 (n = 129) the patients had a mean age of 67 years (range 24-86 years) and consisted of 51 female and 78 male patients. This group contained all of the patients suffering from oesophageal carcinoma (n = 105), and these patients were distributed evenly throughout the antibiotic regime subgroups A, B, and C (see below). For the operations performed, see Table 3. Oesophagectomy was performed by one of three techniques: laparotomy and right thoracotomy, a left thoracophrenotomy approach, or a transhiatal technique [I, 31, 141. All patients undergoing oesophagectomy except a single patient in subgroup A were suffering from oesophageal cancer. In group 2 (n = 97) the patients had a mean age of 61.8 years (range 25-84 years) and consisted of 62 female and 35 male patients. The operations performed were either an antireflux proce-
562 dure, with a Moghissi modified total fundoplication [12] or myotomy for achalasia of the cardia [5]. All study patients in group 1 were randomised at the time of operation to one of three antibiotic regimes. Group A (n = 42) were treated with cefuroxine (CFX) 1.5 g at induction of anaesthesia and then CFX 750 mg b.i.d. for 4 days. Group B (n = 46) were treated with CFX 1.5 g and metronidazole (MTR) 1.0 g at induction of anaesthesia, then CFX 750 mg b.i.d. and MTR 500 mg qds for 4 days. Group C (n = 41) were treated with CFX 1.5 g and MTR 1.O g at the induction of anaesthesia and no further antibiotics unless clinically indicated. Group 2 patients were randomised to one of two antibiotic regimes: group D (n = 47) were treated with CFX 1.5 g on induction of anaesthesia alone, while group E (n = 50) were treated with CFX 1.5 g on induction of anaesthesia, then CFX 750 mg b.i.d. for 2 years. The criteria for infection used in the study were as follows: - Wound sepsis: a discharging surgical wound containing pus, with or without pathogens. _ Pulmonary infection: the presence of clinical (fever, leucocytosis) and radiological signs of pulmonary infection with sputum containing pus, with or without pathogens. - Pleural infection: the presence of pus in the chest (empyema). _ General infection: clinical signs of infection in association with blood cultures yielding pathogens. All patients had routine samples of sputum sent for microbiological examination pre-operatively and on the 3rd, 7th and 11th post-operative days. Wounds were inspected daily for evidence of infection. Any potential episode of infection was investigated using an infection screen consisting of sputum culture, wound swab, urine culture and blood culture. If any other microbiological investigations were considered necessary they were ordered on an individual basis. When an infective complication occurred, administration of CFX was either increased to the treatment dose (CFX 750 mg tds) or restarted at this dosage. MTR was given if still in use as part of the prophylaxis regime, but if it had already been discontinued it was not restarted routinely for infective complications. Treatment with CFX was continued until definitive microbiological results from the infection screen were available, then the antibiotic regime altered accordingly if necessary. Any infective complication occurring during the hospital admission for surgery was considered in this study. All patients were routinely screened with contrast studies on the 4th (group 2) or the 7th post-operative day. Results were analysed using x ’ tests and contingency tables.
Table 1. Organisms identified from the pre-operative sputum samples and the number of cases in which they were isolated” Organisms
isolated
n
Normal flora Haemolytic streptococcus Hemophilus injluenzae Streptococcus pneumoniae Citrobacter sp. Enterobacter sp. Escherichia coli Klebsiella sp. Proteus sp. Pseudomonas sp. Staphylococcus aureus Streptococcus faecalis Candida albicans
192
14 13
a Thirteen patients had more than one organism their pre-operative sputum samples
Table 2. Distribution Antibiotic regime (induction/ post-op.) Group 1 A CFX/CFX B
CFX/CFX MTR/MTR
C
CFX MTR
Group 2 D CFX E
CFX/CFX
of the infective complications Infections ~ n %
identified
from
by group
Site Lungs
Wound
9
21.4
4
3
4
8.6
2
1
18
43.9
13
4
5
10.6
2
3
2
2
Genera1
Pleura
2 1 1
2
CFX, Cefuroxime; MTR, Metronidazole ’ Three patients in group C suffered infection in more than one site
Results
Within the study group of patients there were no instances of oesophageal leakage and no in-hospital death. The pre-operative sputum samples showed potentially pathogenic organisms in 15% of cases. The distribution and frequency of each the isolates from the preoperative sputum samples is shown in Table 1. We could find no correlation between the pathogenic organisms identified from the pre-operative sputum samples and the organisms identified from any subsequent infective complications (P > 0.05). The surgical technique employed for each type of procedure in each of the two study groups for any of the antibiotic prophylaxis regimes made no significant difference to the infective complication rate (P > 0.05). Group 1 patients suffered a total of 31 infective episodes during the study, while group 2 patients suffered 7 infective episodes: the difference was significant (P < 0.05). The most frequently identified infection was pulmonary sepsis: 19 cases in group 1 and 2 cases in group 2. The higher incidence of pulmonary sepsis in
group 1 is a highly significant result (P < 0.001). In association with pulmonary sepsis, three patients in group 1 also developed empyema; in each of these cases the same organism was cultured from both sites. There were 13 wound infections, 8 in group 1 and 5 in group 2; the higher incidence within group 1 is not significant (P > 0.05). A single patient in group 1 suffered from generalised sepsis. In 92% of cases the isolated organism was sensitive to one of the antibiotics employed for prophylaxis. Table 2 shows the incidence of infective complications with respect to the antibiotic subgroups. In group 1 the incidence of infection was significantly higher in groups A (P < 0.05) and C (P c 0.001) than in group B. When pulmonary sepsis and wound infections were considered we found the same significantly higher incidence in groups A (P < 0.05) and C (P < 0.001) than in group B. Significantly more infections occurred in group C than in groups A (P c 0.05) and B (P < 0.05). In Table 3 incidence of infection is considered by type of operation involved as well as by antibiotic subgroup. Patients who underwent oesophagectomy had a sign&
563 Table 3. Distribution of surgery involved
of infective complications
by group and type
Number of patients
Number with infections
26 35 31
6 (23%) 3 (8.5%) 14 (45%)
Group and operation
Group 1 Oesophagectomy A B C Intubation A B C
(n = 14)
Oesophagoplasty A B C Group 2 Antireflux D E Myotomy D E
(n = 92)
8
1
3 3
2
8 8 7
2 1 2
37 41
3 2
10
1 1
(n = 23)
(n = 78)
(n = 19) 9
cantly higher incidence of infective complications than those undergoing intubation (P < 0.001) or oesophagoplasty (P < 0.05). Within the oesophagectomy subgroup there was a significantly higher incidence of infective complications in groups A (P < 0.05) and C (P < 0.001) as compared to group B. In regard to the other procedures performed, the groups are too small for meaningful statistical analysis. It should be noted that in both the oesophagoplasty and the oesophageal intubation groups, infective complications were more frequent in subgroup C. In group 2 we found a significantly higher incidence of infection in subgroup D (P < 0.05) than in subgroup E. In this group the complication rate for the various types of infection was too low for meaningful statistical analysis to be performed. The operative technique involved did not affect the infective complication rate (P > 0.05). Discussion
The use of antibiotic prophylaxis is now considered to be routine in most major surgical procedures on the alimentary tract, and there are numerous studies which support this viewpoint [2, 4, 151, although it should be stressed that antibiotic prophylaxis is never a substitute for good surgical practice and meticulous aseptic technique. So far, no other clinical trials have considered the question of antibiotic prophylaxis for major oesophageal surgery. The study groups reflect the epidemiology of the principal disease types in each of the groups: group 1 contained all patients with carcinoma of the oesophagus and thus the greater number of male patients in the study, while group 2 consisted principally of patients with reflux disease and therefore had the greater number of female patients.
CFX was chosen as our antibiotic for prophylaxis in this study because it has a broad spectrum which would provide adequate prophylaxis against the potential pathogenic organisms, and because of its good penetration into the likely sites of sepsis [9]. In subgroups B and C, MTR was added to enhance the anaerobic cover provided, in view of the spectrum of organisms likely to be encountered [IO]. These suppositions were confirmed by the spectrum of organisms encountered, most of which were sensitive to be antibiotic regime employed. In group 1 we found a considerably higher infective complication rate than in group 2. The incidence of pulmonary sepsis in group 1 was similar to that found in the studies of Fan, Keeling and Galanduik et al. [6,7,8]. The higher infective complication rate within group 1 reflects a number of factors. First, in this group the alimentary tract was opened, and therefore the operative field must be considered to be contaminated. Secondly, the general condition of the patients in group I, a high proportion of whom were suffering from malignant disease, was not as good as in group 2. Lastly, in general patients underwent more extensive surgical procedures, and our results show that the patients undergoing oesophagectomy sustained more infective complications than those undergoing other operative procedures. The conclusion from this is that resectional surgery of the oesophagus is the greatest single risk factor for infective complications, and that the more extensive the operative procedure, the greater the risk of an infective complication. Within group 1 we found a lower incidence of infective complications in subgroup B than in the other two subgroups. Since the subgroups were comparable, the only difference being the antibiotic prophylaxis regime, it must be concluded that the antibiotic prophylaxis given to subgroup B was the most efficacious. This conclusion is true in regard to pulmonary sepsis and wound infection. If the patients undergoing oesophagectomy are considered separately the results are much more marked, with subgroup B clearly having superior prophylaxis to groups A and C. The higher incidence of infective complications within the oesophagectomy subgroup appears to accentuate the superiority of this antibiotic prophylaxis regime. For oesophagoplasty and intubation we did not have sufficient numbers to statistically prove the superiority of the antibiotic prophylaxis regime in subgroup B, although the incidence of infective complications was still lowest in this subgroup. Within group 2 there was a significantly lower incidence of infective complications in subgroup E than in subgroup D. The numbers involved are too low to perform meaningful statistical analysis in regard to the various sites of infection. The operative procedure employed appeared to have no role to play in the frequency of infective complications in group 2.
Conclusions
1. Adequate antibiotic prophylaxis for surgery to the upper gastrointestinal tract when the lumen of a viscus is to be opened is provided by CFX 1.5 g and MTR 1.0 g at
564
induction of anaesthesia, then CFX 750 mg b.i.d. and MTR 500 mg qds for 4 days. 2. Adequate antibiotic prophylaxis for surgery to the upper gastrointestinal tract when the lumen of a viscus is not opened is provided by CFX 1.5 g on induction of anaesthesia, then CFX 750 mg b.i.d. for 2 days. As a resuh of this study, the antibiotic prophylaxis regimes outlined above are now those used in our unit.
7.
8. 9.
IO. Acknowledgement. We would like extend thanks to Dr Meigh and the staff of the microbiology department of Castle Hill Hospital for their assistance in this study.
11.
References
12.
1. Churchill ED, Sweet RH (1942) Transthoracic resection of tumours of the stomach and oesophagus. Ann Surg 115: 897 -917 2. Condon RE, Bartlett JG, Greenlee H, Schulte WJ, Ochi S, Abbe R, Caruana JA, Gordon HE, Horsley JS, Irvan G 3rd, Johnson W, Jordan P Jr, Keitzer WS, Lempeke R, Read RC, Schumer W, Schwartz M, Storm FK, Vetto RM (1983) Efficacy of oral and systemic antibiotics prophylaxis in colorectal surgery. Arch Surg 118: 496 3. Earlam RJ, Cunha-Melo JR (1980) Oesophageal squamous cell carcinoma. 1. A critical review of surgery. Br J Surg 67: 381-390 4. Edmondson HT, Rissing JP (1983) Prophylactic antibiotics in colon surgery. Arch Surg 118: 217 5. Ellis FH Jr, Olsen AM, Schlegel JF, Code CF (1964) Surgical treatment of oesophageal hypermotility disturbances. JAMA 188: 862-866 6. Fan ST, Lau WY, Yip WC, Poon GP, Yeung C, Lam WK, Wong KK (1987) Prediction of post operative pulmonary com-
13.
14. 15.
plications in oesophogastric cancer surgery. Br J Surg 74: 408410 Galanduik S, Hermann RE, Gassman JJ, Cosgrove DM (1986) Cancer of the oesophagus: the Cleveland clinic experience. Ann Surg 203: 101-108 Keeling P, Gillen P, Hennessy TP (1988) Oesophageal resection in the elderly. Ann R Co11 Surg Engl 70: 34-36 Kucers A, Bennet N (1979) Cefamandole, cefoxitin, cefuroxime and ceforanide. In: Kucers A, Bennet N (eds) The use of antibiotics. Heinemann, London, pp 267-293 Kucers A, Bennet N (1979) Metronidazole. In: Kucers A, Bennet N (eds) The use of antibiotics. Heinemann, London, pp 761-781 Lewis I (1946) The surgical treatment of carcinoma of the oesophagus with special reference to growths of the middle third. Br J Surg 34: 18 Moghissi K (1986) Essentials of thoracic and cardiac surgery. Heinemann, London, pp 232-233 Muller JM, Erasmi M, Stelzner M, Zieren U, Pichlmaier H (1990) Surgical therapy of oesophageal carcinoma. Br J Surg 77: 845- 857 Orringer MB, Sloan H (1978) Esophagectomy without thoracotomy. J Thorac Cardiovasc Surg 76: 643 -654 Portnoy J, Kagan E, Gordon PH, Mendelson J (1983) Prophylactic antibiotics in elective colorectal surgery. Dis Colon Rectum 26: 310
Mr D. A. C. Sharpe, MB, FRCS North Humberside Cardiothoracic Centre Castle Hill Hospital Castle Road Cottingham, North Humberside HU16 5JQ United Kingdom
