Endoscopic ultrasound (EUS) with fine-needle aspiration (FNA) is universally used for the investigation and diagnosis of pancreatic cystic lesions (PCL). Infectious complications (i.e. fever, infected cyst or bacteremia) following EUS-FNA are uncommon, with a reported incidence of 0-6% in various studies [1-6]. Furthermore, prospective data from studies that used antibiotic prophylaxis suggest a very low risk (0-1.4%) of infectious complications [7-9]. However, whilst the incidence is low, post-FNA infection can result in readmission to hospital and adverse patient experience.
Although antibiotic prophylaxis to reduce the risk of infection is recommended by both the American Society of Gastrointestinal Endoscopy and the European Society of Gastrointestinal Endoscopy, there is no consensus on the optimal regimen or route of administration [1,10]. Previous studies reported on the use of single-dose intravenous (IV) regimens, such as ciprofloxacin, or piperacillin/tazobactam, while commonly used 3-5 day oral regimens include ciprofloxacin or amoxicillin/clavulanic acid [7,11]. Potential advantages of a single-dose IV antibiotic over a prolonged oral regimen include simplicity, guaranteed delivery and potentially fewer antibiotic related adverse events; however, there are only limited data to support this.
Ceftriaxone is a third-generation semisynthetic cephalosporin with a long half-life. It is administered IV or intramuscularly, with an overall excellent safety profile, and has a broad spectrum of activity against Gram-positive, Gram-negative aerobic, and some anaerobic bacteria [12,13]. Ceftriaxone is well-established as the drug of choice for surgical prophylaxis [14,15].
We herein aimed to establish the safety and efficacy of a single-dose intraprocedural antibiotic (1 g IV ceftriaxone) in preventing infectious complications following EUS-FNA of PCL.
Patients and methods
We conducted a single-center retrospective analysis of the safety and efficacy of single-dose IV antibiotics during EUS-FNA of PCLs. The study was approved by the local human research and ethics committee (HREC LNRSSA/15/WMEAD/121).
Patients who underwent ambulatory EUS-FNA of PCLs at our hospital between March 2006 and December 2015 were identified through an endoscopy reporting database program (ProVation MD, version 5 - ProVation Medical Inc. USA; or Endoscribe - Health Communication Network, Australia).
Westmead Hospital is a tertiary referral center for the Western Sydney local health district, with an annual EUS case volume of over 600. Two consultant gastroenterologists with extensive EUS experience, or their advanced endoscopy fellows under direct supervision, performed all EUS procedures. Procedures were performed under conscious sedation using midazolam, fentanyl, and propofol. Standard EUS-FNA technique was used in all cases. The choice of FNA needle size was at the discretion of the endoscopist: generally, a 19 G or 22 G needle was used for a trans-gastric approach and a 22 G or 25 G for a trans-duodenal approach. Whenever possible cysts were drained to dryness. The decision to administer periprocedural antibiotics (1 g IV ceftriaxone) for non-penicillin allergic patients, or to prescribe oral antibiotics (3-5 days of amoxicillin/clavulanic acid or ciprofloxacin) on discharge from the endoscopy unit was at the discretion of the consultant performing the procedure. Patients were monitored for 1-2 h following the procedure prior to discharge. Analysis of fluid amylase and carcinoembryonic antigen were performed in the hospital pathology lab. Cytology was reported by a specialist cytologist.
Data collection included patient demographics, procedural data and outcome. We reviewed the patient’s endoscopy report, their computerized medical records and any presentation to a hospital in our district within 30 days of the procedure. In our health district, computerized medical record software (Powerchart), which includes procedure reports, admission and discharge letters, imaging studies and all laboratory data, has been available since 2006.
An infectious complication was defined as fever/rigors, or bacteremia, accompanied by clinical symptoms such as abdominal pain, nausea and vomiting, or by radiological or laboratory evidence of an abdominal infection, all within 30 days of the procedure. Pancreatitis was defined as a ≥3-fold elevation in amylase or lipase, or typical findings on computed tomography accompanied by an appropriate clinical scenario.
Continuous data are presented as mean + standard deviation or median + interquartile range (IQR). Categorical data are presented as frequencies. The association between categorical variables was assessed using the chi-square or the Monte Carlo exact test, as needed.
One hundred eighty-four patients (mean age 63 years, 60% female), underwent EUS+FNA of 204 PCLs (median size 20 mm, IQR 16-30) between March 2006 and December 2015. Single-dose IV ceftriaxone was given in 146 procedures, oral antibiotics for 3-5 days were given in 23 procedures and in 35 procedures periprocedural antibiotics were inadvertently omitted (Fig. 1).
Figure 1 Study flowchart
The main indications for the procedure included evaluation of newly diagnosed PCL (84%), or surveillance of known PCL (16%). Cyst types were serous (23%), mucinous (14%), intraductal papillary mucinous neoplasms (40%), cystic tumors (17%), or undetermined (6%). Cysts were located in the pancreatic head (31%), neck (12%), body (38%), or tail (12%). Morphologically, cysts were also classified as simple (61%), oligocystic (25%), polycystic (9%), or suspected cystic tumor (5%). Fluid type was recorded as clear/watery (62%), thick/viscous (30%), turbid (5%), or pus (3%).
EUS+FNA was attempted in 204 PCLs. Fluid was successfully aspirated in 191/204 (94%) of the cases. Most commonly, a 22 G needle was used (72%) followed by 19 G (17%) and 25 G (11%). An average of 1.3 passes were made (range 1-5). A summary of the descriptive data is presented in Table 1.
Table 1 Patient lesions and procedural characteristics
A total of 6 (3%) procedure-related complications were recorded in patients receiving antibiotics. Three patients had post-procedure pancreatitis (2 patients that received an oral regimen and 1 patient who received IV ceftriaxone), and 3 patients had postoperative pain that was resolved in all cases with conservative management. No infectious adverse events were encountered and no antibiotic-related adverse events were recorded in patients receiving periprocedural antibiotics. In the group of patients who did not receive antibiotics (n=35), there was one case (3%) of an infectious complication. Table 2 summarizes the complications and outcomes.
Table 2 Complications in patients receiving intravenous ceftriaxone (n=146)
Since our study spanned nearly a decade of clinical practice, we also analyzed the results according to years. We compared the results from procedures performed up to and including 2010 to those performed between 2011 and 2015. We found no differences in the indications, percentage of procedures where antibiotics were administered, the type of antibiotics given and the rate of complications (Table 3). We also found no differences in the study endpoints when comparing different cyst types, although the numbers for each group in this comparison were very small and hence it is difficult to draw conclusions.
Table 3 Comparison between procedures performed up to and including 2010 and those performed between 2011 and 2015
EUS+FNA of PCLs has an overall favorable safety profile. However, compared to EUS+FNA of solid lesions, it appears to have a higher risk of infectious complications [1,10]. Current data demonstrate a low risk of infectious complications; however, most studies are retrospective and differ in design, patient characteristics and the use of antibiotics.
In a large retrospective study by Lee et al , 603 FNA were performed and only 1 patient developed an infectious complication. Antibiotics (fluoroquinolones, ampicillin, vancomycin, or gentamicin) were given at the discretion of the endoscopists and usually in cases of large cysts or incomplete drainage. However, this was not associated with a reduction in infectious complications. In a more recent study, Guarenir et al  reported on 1 case of cyst infection and 3 cases of antibiotic-related adverse events in a cohort of 88 patients who were given periprocedural antibiotics. In this study, multiple antibiotic regimens were used (levofloxacin in 78 cases, ciprofloxacin plus metronidazole in 3 cases, vancomycin plus gentamicin in 2 cases, piperacillin-tazobactam in 2 cases, and single-dose amoxicillin-clavulanate plus metronidazole, ampicillin, or doxycycline in 3 cases).
There are few prospective studies reporting on infectious complications in patients following EUS-FNA of PCLs. Barawi et al  reported no infectious complications in EUS-FNA of 108 lesions. In a study by Tarantino et al , 298 patients received a single dose of piperacillin/tazobactam followed by 3-5 days of oral antibiotics). Four patients (1.3%) developed fever following the procedure. Finally, Marinos et al  reported no infectious complications in a cohort of 85 EUS-FNA cases that were given a single IV dose of piperacillin/tazobactam or ciprofloxacin. While the results of these studies are encouraging, complex regimens involving parenteral antibiotics hours in advance of the procedure or oral courses after the procedure increase the complexity of the procedure and may result in non-adherence.
IV ceftriaxone is well-established and frequently used as the drug of choice for surgical prophylaxis [14,15]. It has an excellent safety profile and can be administered as a single IV dose during the procedure, making it an attractive alternative to more complex parenteral and oral regimens previously described. It is also less likely to contribute to the possible future development of complex antibiotic resistance. Although our study was not powered to perform a comparison of complications between the IV ceftriaxone regimen and the oral regimen, or the cases, which did not receive any antibiotics, encouragingly we had no infectious complications and no antibiotic-associated adverse events.
This study has a number of limitations. A retrospective analysis is prone to biases. The patients referred to our endoscopy unit for EUS are generally from our local health district. On discharge following EUS-FNA, all patients are routinely advised to contact us and seek medical treatment if they feel unwell in the days following the procedure. They are provided with a copy of the procedure report, which contains instructions and a telephone number for contacting our center in the event of any complications or re-presentation to medical services. We recognize that theoretically this may occur outside our local health district or in private rooms; however, we believe it is not very likely. We relied on medical records to detect complications following EUS-FNA. We elected not to contact the patients directly, since we believed there would be significant recall bias. Our sample size was not large enough to perform any comparative analysis. However, since the therapeutic intervention reported here has not been previously reported to our knowledge and the body of evidence in this area is lacking, we believe our results are important and will present physicians performing EUS-FNA with an additional, non-inferior alternative for the prevention of infectious complications.
In conclusion, EUS-FNA of pancreatic cystic lesions is safe. A single dose of intraprocedural IV ceftriaxone is a convenient, safe and effective intervention to prevent infectious complications following FNA. These results can serve as a platform to design a multicenter randomized control trial that can further investigate the role of antibiotic prophylaxis during EUS-FNA of PCLs.
What is already known:
Infectious complications following endoscopic ultrasound fine-needle aspiration (EUS-FNA) are uncommon, with a reported incidence of 0-6%
Both the American and the European Societies of Gastrointestinal Endoscopy recommend antibiotic prophylaxis following EUS-FNA of pancreatic cystic lesions to reduce the risk of infection
There is no consensus on the optimal antibiotic regimen or route of administration
Previous reports used complex antibiotic regimens
What the new findings are:
In this retrospective analysis, single-dose IV ceftriaxone was effective for the prevention of infectious complications following EUS-FNA of pancreatic cysts
There were no antibiotic-associated adverse events with this regimen
This offers clinicians a simple and safe alternative to more complex and prolonged antibiotic regimens