Endoscopic ultrasonography (EUS) is an established modality for the diagnosis of esophagogastric, mediastinal, pancreatobiliary, and pelvic lesions. However, it is mainly used for elective procedures in outpatients and occasionally inpatients who are at low risk for complications from moderate or deep sedation. While diagnostic and therapeutic endoscopy procedures such as PEG placement or evaluation of GI bleeding are routinely performed at the bedside in intensive care units (ICUs), the use of EUS for this indication has been described in only two small case series [
This study was approved by the Institutional Review Board at Clarian Health Partners/Indiana University Hospital. Our prospectively updated endoscopy database was queried to identify all patients who underwent a bedside EUS procedure in an ICU at our hospital between January 2004 and August 2009.
Patients in this cohort were deemed by their treating physicians to require ICU care and also an EUS procedure for evaluation of various pancreaticobiliary, thoracic, or pelvic disorders. To complete these procedures, patients were not deemed stable for transport to our endoscopy unit and therefore required a bedside procedure.
Sedation for these procedures varied depending on the presence or absence of previous endotracheal intubation. If the patient was intubated, intravenous sedatives or narcotics already being administered to the patient were given for the procedure either by intermittent boluses or a temporary increase in the maintenance infusion rate. Alternatively, intermittent boluses of propofol were used. If the patient was not intubated, then sedation was accomplished by intermittent boluses of benzodiazepines and/or opioids. Appropriate cardiorespiratory monitoring was used for all patients. Informed consent was obtained from the patient, patients’ relative or power of attorney in all cases.
All procedures were performed by or under the supervision of one of five experienced attending endosonographers. If a radial echoendoscope was required, examination was initiated with an Olympus GF-UE160-AL5 radial echoendoscope (Olympus America, Inc., Center Valley, PA, USA). Curvilinear array endosonography was performed using an Olympus GF-UC30P or Olympus GF-UC140P-AL5 (Olympus America, Inc., Center Valley, PA, USA) echoendoscope. EUS-FNA was performed at the discretion of the endosonographer and if performed was obtained using a 19 or a 22-gauge EUSN-1, EUSN-2, EUSN-3, or Echotip Ultra needle (Cook Medical Inc., Winston-Salem, NC, USA) or EZ-Shot needle (Olympus America, Inc., Center Valley, PA, USA). Doppler examination was used to ensure the absence of intervening vascular structures along the anticipated needle path. Depending on the amount of blood anticipated during tissue sampling, full, partial, or no suction was applied. When a cytopathologist and portable microscope were available on site for preliminary diagnostic interpretations and assessment of specimen adequacy, FNA samples were expressed onto a glass slide and two smear preparations were made. One slide preparation was air-dried and stained with a modified Giemsa stain for on-site review, while the other slide was alcohol-fixed and stained by the Papanicolaou method. In the absence of on-site pathology review, samples were only alcohol-fixed for future staining and interpretation. EUS-FNA was repeated until a definitive diagnosis was made or the endosonographers felt that further sampling would not likely increase yield. Additional FNA passes were made when either (1) a cellblock preparation was potentially required for immunostains or (2) microbiology stains and cultures were required for potential sites of infections. Definitive cytopathologic diagnoses were given only after complete staining and subsequent final interpretation was provided. Hospital records and endoscopy charts were reviewed after the procedure to assess for any short-term complications and clinical followup.
For study purposes, EUS was considered to have a significant impact if a new diagnosis was established and/or the findings altered subsequent management. EUS was not considered to have an impact if a diagnosis was incorrectly interpreted or missed. When available, the final clinical diagnosis for each patient was made by autopsy or surgery. In the absence of either, the results of EUS-FNA and clinical followup were used.
Fifteen patients (9 male; mean age:
Indications for EUS were mediastinal mass or adenopathy (
Bedside EUS was technically successful and completed in all patients without complications. Twelve (80%) patients underwent EUS-FNA (median 4 passes; range 2–7) and a cytopathologist was available on site for preliminary interpretations in 8 of 12 (66.7%). The findings and results of the EUS exams are summarized in Table
EUS findings, final diagnoses, and outcomes of the patients.
Patient | Age/gender | Indication for EUS | EUS finding | Number of passes (FNA) | On-site cytologist available | Cytological finding | Final diagnosis | Impact of EUS on management | Patient’s outcome |
---|---|---|---|---|---|---|---|---|---|
1 | 74/F | Suspected pancreatic mass on CT | Solid-cystic mass at BOP | 3 | Yes | Acute inflammationa | Necrotizing pancreatitis & pancreatic abscess | Yes | Died* |
2 | 66/M | Suspected pancreatic mass on CT | Heterogenous mass in retroperitoneum | 4 | Yes | Suspicious for lymphoma | Systemic Amyloidosis & hemorrhagic pancreatic pseudocyst | No | Died† |
3 | 50/M | Suspected CBD stone | No CBD stone | NA | N/A | NA | UTI & Sepsis | Yes | Alive |
4 | 50/F | Suspected gastric mass | Focal thickening of gastric mucosa (no tumor) | 3 | No | Benign epithelium | Severe Pancreatitisb | Yes | Died† |
5 | 72/M | Suspected extrahepatic cholestasis | No CBD tumor or stone | NA | N/A | NA | Liver failurec | Yes | Died* |
6 | 76/M | Mediastinal LN on CT | Subcarinal LN | 5 | No | Reactive LN | Severe pneumonia | Yes | Died* |
7 | 54/M | Evaluating gastric varices | Thrombosed gastric varices | NA | N/A | NA | Gastric variceal bleedingd | Yes | Alive |
8 | 52/M | Suspected pelvic abscess | Post-surgical cyst | 2 | No | Hypocellular samplee | Hemorrhagic pancreatitis & sepsis | Yes | Died† |
9 | 61/F | Enlarging pancreatic mass | Large portahepatis LNf | 3 | No | Reactive LN | Respiratory failure due to pneumonia | Yesg | Alive |
10 | 37/F | Mediastinal mass on CT | Mediastinal mass | 3 | Yes | Nonsmall cell carcinoma | NSCLC with mediastinal involvement | Yes | Died* |
11 | 43/M | Mediastinal adenopathy on CT | Paraesophageal LN | 6 | Yes | Adenocarcinoma | Metastatic adenocarcinoma | Yes | Died* |
12 | 58/M | Mediastinal mass on CT | Mediastinal mass | 4 | Yes | SCC | NSCLC | Yes | Died† |
13 | 76/M | Mediastinal adenopathy on CT | Pleural effusion; celiac & perigastric LN | 7 | Yes | Non-Hodgkin’s lymphoma | Non-Hodgkin’s lymphoma | Yes | Died† |
14 | 28/F | Mediastinal adenopathy on CT | Mediastinal mass | 5 | Yes | Necrosis | Histoplasmosis | No | Alive |
15 | 67/F | Mediastinal mass on CT | LN in aortopulmonary window | 4 | Yes | Small cell carcinoma | Small cell lung cancer | Yes | Died† |
Abbreviations: BOP: body of pancreas; CBD: common bile duct; NA: not applicable; UTI: urinary tract infection; LN: lymph node; NSCLC: nonsmall cell carcinoma; SCC: squamous cell carcinoma.
*Died a few months after discharge from ICU.
†Died while was in ICU.
aEnterococcus grew on culture of fine needle aspirate demonstrating pancreas abscess.
bPatient’s final diagnosis was severe pancreatitis of allograft pancreas, superior mesenteric vein and superior mesenteric artery thromboses, and abdominal compartment syndrome.
cLiver failure due to alcoholic cirrhosis.
dPatient had gastric variceal bleeding due to alcoholic cirrhosis. He underwent EUS for possible EUS-guided injection of cyanoacrylate, but the varices were found thrombosed from the cyanoacrylate injection one week before.
eCulture of fine needle aspirate was negative. Autopsy showed no pelvic abscess.
fVery large portahepatis LN with heterogeneous echotexture suggestive of recent bleeding.
gEUS-FNA had significant impact on the patient management through excluding malignancy.
EUS-FNA provided an initial diagnosis of a malignant mediastinal mass/lymph node in five including nonsmall cell lung cancer (NSCLC) in two, small cell lung cancer (SCLC) in one, metastatic adenocarcinoma in one, and non-Hodgkin’s lymphoma (NHL) in one. Additionally, EUS-FNA diagnosed a pancreas abscess (
EUS failed to establish a correct diagnosis in two patients. In one (patient 2, Table
Six (40%) patients died during their ICU admission and the remaining nine were discharged. Five of these died within three months of hospitalization while four remained alive a median 321 days (range: 209–706 days) after discharge. No short-term complications were noted during or immediately after EUS.
There are limited data previously published in two small series that describe the utility of bedside EUS in critically ill patients. Fritscher-Ravens et al. [
Varadarajulu et al. [
In the current series, we found that bedside EUS had a clinical impact in 13 of 15 critically ill patients evaluated. A new diagnosis was established in 6 cases and precluded the need for further studies or intervention in seven others. We found that the most common clinical impact made by EUS was providing a tissue diagnosis of malignancy. However, as the other two reported series [
The current study represents the largest series to date detailing the use of bedside EUS and confirms its utility and impact in this patient population. Nevertheless our study is limited by its retrospective design and limited followup in some patients. Further prospective studies involving larger numbers of patients are needed to further define the role of EUS in critically ill patients.
In conclusion, bedside EUS in ICU patients is safe and feasible. In the current series, it had a significant impact on the management of the majority of patients evaluated. Therefore, endosonographers should consider using EUS in ICU patients when clinically indicated. Given the expanding field of EUS, we postulate that therapeutic procedures such as pancreatic necrosectomy [
All authors disclose no financial relationships relevant to this publication.
M. Mohamadnejad and J. Dewitt contributed in conception, design, obtaining IRB approval, conducting the study, collecting and analyzing the data, and writing the paper. J. LeBlanc, S. Sherman, M. Al-Haddad, L. McHenry, and G. Cote contributed in conception, design, and conducting the study. All the authors approved the final version of the paper.