Abstract
Abstract
Background:
Esophageal cancer is a serious diagnosis that has a relative incidence of 4/100,000 inhabitants in the Czech Republic. This disorder is managed predominantly by surgery. The steps to improving the outcome of treatment include a multifactorial approach. The role of operative technique in improving outcomes seems to have reached its limits. However, antibiotic prophylaxis and the treatment of complicating bacterial infections continue to play important roles.
Methods:
A total of 85 patients with strictly defined antibiotic prophylaxis during surgical esophagectomy were included in our study. Bacterial strains were isolated from the patient's clinical materials after operation; only one strain from each patient, the first to be isolated, was tested for antibiotic sensitivity.
Results:
Infectious complications were observed in 15.3% of patients and the mortality rate from infectious complications reached 30.8%. The most frequently documented complicated infection was pneumonia (69.2%) and the most frequent pathogens were enteric bacteria (56.5%). Some bacterial strains producing extended-spectrum beta-lactamases and AmpC beta-lactamases were found.
Conclusions:
The infections in our patient set were of endogenous origin. In cases of pneumonia, it is appropriate to begin with antibiotics effective against enteric bacteria and Pseudomonas aeruginosa.
Esophageal cancer usually is treated surgically. Currently, mini-invasive procedures are preferred. Transhiatal laparoscopic esophageal extirpation was introduced in the Department of Surgery I at the University Hospital Olomouc, Czech Republic, in 2001. In patients in whom the esophagus can be removed without thoracotomy, which is much better tolerated by the patients, observation of the extirpation may be ensured using the transhiatal laparoscopic procedure. For proximal tumors located in the thoracic esophagus, esophagectomy is performed via a thoracic approach, either thoracoscopically or by thoracotomy.
A complex of measures, including nutritional support, immunologic management, and other pre-operative preparation, along with modern anesthesia and peri-operative and post-operative care, has improved the immediate results of surgical treatment substantially in esophageal cancer. However, the long-term results remain poor, and the operative technique cannot affect them to a major degree. A condition for reaching good long-term results is elimination of peri-operative and post-operative complications. To prevent such complications, it is necessary that antibiotic prophylaxis be ensured. In cases of complicating bacterial infection, adequate antibiotic treatment arising from knowledge of the most frequent bacterial pathogens and their resistance to antimicrobial preparations is required. The aim of the present work was to analyze infectious complications after esophagectomy and to determine appropriate antibiotic therapy.
Patients and Methods
The data in this study were collected retrospectively, and in every case, we obtained the patient's informed consent for the study. All patients received neoadjuvant chemotherapy and radiotherapy, and the operation was performed 6–8 weeks after its completion. Prior to the operation, a chest radiograph was performed, and none of the patients showed signs of mediastinitis.
All patients underwent a nutritional evaluation prior to the operation, which also consisted of laboratory workup including measurement of the serum albumin concentration. Nutritionally compromised patients were prepared before the operation or, in cases of very poor nutritional score, were not subjected to operation. Pre-operatively, a biluminal nasogastric tube was introduced to ensure nourishment post-operatively.
Prior to the operation, all patients provided a urine sample for analysis to rule out urinary infection. If a urinary infection was confirmed, it was treated, and the operation was performed once the infection had resolved.
If the patient exhibited any signs of infection, he or she was examined thoroughly, and any infection was treated before the operation. The patient had a thorough laboratory workup performed prior to the operation, including glycemic measurement. To prevent infection peri-operatively, the patient received prophylactic antibiotics. Two layers of antiseptic solution were applied to the skin of the operative site. Normal temperature was maintained in the operating room.
Surgical procedure
If the procedure is performed laparoscopically and transhiatally, it involves dissection of the diaphragmatic crura and isolation of the distal, or thoraco-abdominal, part of the esophagus. Under videoscopic control, one can pass safely around the tumor in the mediastinum. Paraesophageal lymph nodes may be identified, especially in the esophageal–aortic space. If they are found, they are collected along with the specimen. This method may be performed relatively easily up to the bifurcation of the trachea. Then, from a left cervical incision, the cervical and upper thoracic esophagus is dissected so that the esophagus is almost completely liberated. It is transected in the cervical region and extirpated. Laparoscopically, or from a small laparotomy, liberation of the stomach is completed with preservation of the vascular arcade of the right gastroepiploic vessels. Gastric tubularization along the greater curvature is performed with subsequent pyloromyotomy. The transponate is then pulled up to the cervical area, where the esophagogastric anastomosis is constructed.
Liberation of extensive or proximally located tumors requires a thoracic approach. The esophagus with the tumor is extirpated either by thoracotomy or, if possible, thoracoscopically, which is advantageous because this approach is less burdensome for the patient. The reconstruction phase does not differ. It consists of attaching the cervical esophagus to the tubularized stomach; an exception may be anastomosis construction in the thoracic cavity.
Antibiotic prophylaxis
In all patients, the surgical procedure was performed under antibiotic prophylaxis using amoxicillin/clavulanic acid in three doses of 1.2 g intravenously. The first dose was administered at induction of anesthesia, the second dose 2 h later, and the third dose 4 h after the second dose.
Definitions of infection complications
Pneumonia was defined as the presence of new or progressive infiltrates on chest radiographs plus at least two other signs of respiratory tract infection: Temperature >38°C, purulent sputum, leukocytosis >10×103/mm3 or leukopenia<4×03/mm3, signs of inflammation on auscultation, cough, or respiratory insufficiency with an oxygenation index (PaO2:FIO2) ≤300.
Blood stream infection was characterized by bacteremia and the following clinical signs: Tremor, temperature >38°C (hyperthermia) or <35.6°C (hypothermia), pulse frequency >90/min, tachypnea (>20 breaths/min), left-shifted leukocytosis (>12,000/mm3, >10% immature forms) or leukopenia (<4,000/mm3), and inflammatory markers (e.g., C-reactive protein [CRP], interleukins, procalcitonin) [2]. Surgical site infections (SSIs) were defined in accordance with the U.S. Centers for Disease Control and Prevention rules [3]. Superficial incisional, deep incisional, and organ/space SSI were diagnosed as defined in the same document [3]. Infectious complications were detected by analysis of drain secretions, sputum cultures, temperature, biochemical methods, CRP, leukocytes, and clinical condition.
Isolation and identification of bacterial and mycotic pathogens
Bacterial strains were isolated from clinical materials (sputum, pus, blood, exudates) and identified by standard microbiological methods and the Phoenix automated system (Becton Dickinson, Heidelberg, Germany). Strain selection was realized so that from each patient, only one strain was chosen: The one isolated first. The isolates were preserved at −70°C using ITEST Kryobanka B (ITEST plus s.r.o.; Hradec Králové, Czech Republic) until testing. Yeasts were identified on the basis of cultivation characteristics and biochemical testing.
Antibiotic sensitivity testing
Sensitivity to various antibiotics was estimated by a standard dilution micromethod [4,5]. Reference strains—Escherichia coli American Type Culture Collection (ATCC) 25922, E. coli ATCC 35218, and Pseudomonas aeruginosa ATCC 27853 (Rockville, MD)—were used for protocol quality controls. A modified double-disk synergy test was used for phenotype detection of extended-spectrum beta-lactamases (ESBLs) [6]. Phenotypic determination of ESBL production was confirmed by polymerase chain reaction (PCR) amplification of blaTEM, blaSHV, and blaCTX-M, which encode the most common types of these beta-lactamases [7–9]. AmpC beta-lactamases were determined by multiplex PCR capable of differentiation of various types of these enzymes [10].
Results
Eighty-five patients underwent radical operation for esophageal cancer from 2006 to 2009. Of these patients, 71 did not require thoracotomy; in 14 cases, the esophagus was liberated via the thoracic approach: eight thoracoscopically and six via thoracotomy. The patient set included 72 men with an average age of 58.6 years (range 34–77 years) and 13 women with an average age of 63.5 years (range 45–78 years).
Infectious complications were observed in 13 patients (15.3%) (Table 1). Death associated with complicating bacterial infection was observed in four patients (4.7% of the entire series and 30.8% of those with infections). The infection documented most frequently was pneumonia (69.2%), which in all cases were ventilator-associated (VAP). The overall rate of VAP in the Department of Surgical Intensive Care was 18.5%. The duration of mechanical ventilation ranged from 12 to 24 h after the operation. The frequency of SSI was 23.1%, and every SSI was superficial. The primary reason for incisional infections was an anastomotic leak. Blood stream infection was observed in one patient and was related to a central line. Urinary tract infections were not observed.
The most common causative agents of infectious complications were enteric bacteria (Escherichia coli, Klebsiella pneumoniae, and Enterobacter cloacae strains). Their etiologic role was established in 57.1% of the patients with pneumonia. In cases of SSI (mainly mediastinitis), enteric bacteria were the causative agent in 62.5%. Polymicrobial infections were seen in three patients with pneumonia and in two patients with SSI. The anaerobic bacterium Bacteroides fragilis was the causative agent in one case of SSI. Death associated with infectious complications was recorded in four patients (Table 2).
The sensitivities of gram-negative pathogens to various antibiotics are presented in Table 3. Production of ESBL type CTX-M was observed in one strain of E. coli and one strain of K. pneumoniae. These strains acted concomitantly as the bacterial pathogens responsible for the fatal mediastinitis in Patient 3. In one strain of E. cloacae, which together with a multi-resistant strain of P. aeruginosa caused pneumonia and subsequent death, production of AmpC beta-lactamase was observed, which undermines resistance to extended-spectrum penicillins and cephalosporins, including their combinations with inhibitors of bacterial beta-lactamases.
AMI=amikacin; AMS=ampicillin-sulbactam; CIP=ciprofloxacin; COL=colistin; CPM=cefepime; CRX=cefuroxime; CTX=cefotaxime; CTZ=ceftazidime; GEN=gentamicin; MER=meropenem; PPT=piperacillin-tazobactam; TIG=tigecycline.
Based on the results of sensitivity testing of the gram-negative pathogens isolated, taking into consideration the small number of isolates within individual bacterial species, it may be stated that the most effective antibiotics were meropenem, colistin, tigecycline, and amikacin. The Staphylococcus aureus isolate was an oxacillin/methicillin-sensitive strain. The Enterococcus faecium strain was resistant to ampicillin and sensitive to vancomycin and teicoplanin.
Discussion
In our patient set, infectious complications were observed in 13 patients (15%). There is a difference in the percentage of complications in patients operated on classically, where the percentage reached 45%, in comparison with the mini-invasive procedures, where the frequency of complicating bacterial infections was 19% [11]. Braghetto et al. say that the most common infectious complications in patients undergoing esophagectomy are pulmonary [12]. In our set of patients, pneumonia represented 69% of all infectious complications. In their study, Akutsu et al. present a 36% incidence of post-operative pneumonia in patients undergoing esophagectomy [13].
From the determination of the bacterial and mycotic causative agents of all infectious complications in our patients, it is apparent that the most frequent pathogens (74%) were enterobacteria (E. coli, K. pneumoniae, and E. cloacae strains) and non-fermenting gram-negative bacteria (P. aeruginosa, Burkholderia cepacia, and Acinetobacter baumannii).
It is important to note that of the enteric bacteria, two strains (E. coli and K. pneumoniae) were shown to produce ESBLs, and one strain, E. cloacae, produced AmpC enzymes. Production of these enzymes promotes resistance to multiple antimicrobial drugs and limits the possibilities for adequate antibiotic therapy. It is probable that such multi-resistant bacterial strains contributed to the lethal course of complicating bacterial infection, as antibiotic prophylaxis and subsequent initial antibiotic therapy (piperacillin-tazobactam alone or in combination with gentamicin) was ineffective against these pathogens. Thus, the results indicate the negative connotation significance of the development of bacterial resistance, thereby leading to the failure of antibiotic therapy and associated higher morbidity.
Furthermore, it may be assumed that the infectious complications in our patient set were endogenous and that the bacterial or mycotic etiologic agents came from location-specific microflora, in the case of pneumonia from the upper respiratory tract. However, it is necessary to take into account the probable secondary colonization by multi-drug-resistant bacteria, which is supported by the presence of ESBL-positive strains of E. coli and K. pneumoniae, as well as the AmpC-positive isolate of Ent. cloacae. This presents a question regarding the benefit of testing colonization of the upper respiratory tract preoperatively and, if multi-resistant bacteria are found, changing the antibiotic prophylactic regime according to the results of antibiotic sensitivity or resistance testing.
It is apparent that in cases of complicating pneumonias in patients who have undergone esophagectomy, it is appropriate to begin antibiotic therapy effective against enteric bacteria, including strains producing ESBLs and P. aeruginosa. This specification is best met by carbapenems (for example, imipenem-cilastation or meropenem), tigecycline in combination with an aminoglycoside, and perhaps piperacillin-tazobactam. In cases of SSI, the above antibiotics may also be recommended, possibly in combination with clindamycin or metronidazole.
Footnotes
Acknowledgment
This work was supported by the grant of Ministry of Education, Youth, and Sports MSM6198959223 and by the internal grant of the Palacky University Olomouc LF_2001_002.
Author Disclosure Statement
No conflicting financial interests exist.