Endoscopy in neutropenic and/or thrombocytopenic patients

Table 1 Comparison of United States and British guidelines for endoscopy in thrombocytopenic and neutropenic patients

	US guidelines	British guidelines
Thrombocytopenia and endoscopy	ASGE: Acknowledge limited data. Platelet threshold 20000/mm3 for diagnostic endoscopy; 50000/mm3 if biopsies performed	BSG: Ensure platelet support is available before endoscopic intervention when platelet count is < 50000-80000/mm3
Neutropenia and endoscopy	ASGE: Recommend considering antibiotic in immunosuppressed patients undergoing a high-risk procedure	BSG: Recommend antibiotic prophylaxis for ANC < 500/mm3 and undergoing a high risk procedure (based on risk of bacteremia in immunocompetent patients)

ASGE: American Society for Gastrointestinal Endoscopy; BSG: British Society of Gastroenterology.

Table 2 Study design and characteristics of patients and endoscopies

Study	Design	Patient and endoscopic characteristics
Buderus et al[19] (2012)	Retrospective 1995-2004	38 pediatric cancer patients with various GI complaints
		40 diagnostic endoscopies, 7 follow-up endoscopies, 10 therapeutic endoscopies
		Diagnostic yield 82.5%: Gastritis, esophagitis, duodenitis, colitis, Mallory-Weiss tears, ulcer
Chu et al[17] (1983)	Retrospective 1978-1979	133 cancer patients with thrombocytopenia and overt GI bleed
		187 diagnostic endoscopies, no therapeutic endoscopies
		Diagnostic yield 92% for upper, 60% for lower exam: Unifocal and multifocal lesions in majority; rare diffuse bleeding
Gorschlüter et al[20] (2008)	Retrospective 1993-2005	104 acute leukemia patients after myelosuppressive chemotherapy
		131 primary endoscopies, 40 follow-up endoscopies; includes 16 therapeutic interventions and 5 ERCPs (2 for jaundice, 2 for suspicion of cholecystitis, 1 for suspicion of cholangitis)
		Diagnostic yield 91% for upper, 70% for lower exam: esophagitis, gastric erosions, hiatal hernia, gastritis
Kaur et al[22] (1996)	Retrospective 1986-1993	43 post-bone marrow transplant patients with overt GI bleed
		31 endoscopies total: 26 EGD, 5 colonoscopy; 2 endoscopies required hemostasis
		Diagnostic yield 100% for upper, 80% for lower exam: Diffuse esophagitis, gastritis, or duodenitis in upper exam; 2 ulcers, 1 colitis, 1 tumor recurrence in lower exam
Kaur et al[23] (2013)	Retrospective 2007-2010	11 pediatric patient requiring PEG placement in anticipation of BMT (BMT group) compared with 30 patients requiring PEG placement for other indications (comparison group)
Khan et al[24] (2006)	Retrospective 1995-2002	191 pediatric patients who underwent hematopoietic stem cell transplantation
		198 EGDs, 220 lower endoscopies. All diagnostic endoscopies for GI complaints, mostly for nausea, vomiting, and non-bloody diarrhea.
		Diagnostic yield 32% for upper, 16% for lower exam:
		Mucosal abnormalities most common
		Acute GVHD in 14% on histological exam
		Non-GVHD histological evidence of inflammation in 24%
Park et al[21] (2010)	Retrospective 2002-2007	32 patients with aplastic anemia and overt GI bleed, each evaluated by endoscopy, 3 of which required therapeutic intervention
		Diagnostic yield 66%: bleeding sites in esophagus, stomach, duodenum, small intestine, large intestine
Ross et al[25] (2008)	Retrospective 2002-2006	112 patients with simultaneous upper and lower endoscopic procedures following hematopoietic stem cell transplant. All diagnostic endoscopies for GI symptoms
		Diagnostic yield: GVHD diagnosed in 81% of patients
Schulenburg et al[26] (2004)	Prospective cohort 1996-2001	42 post-allogeneic stem cell transplant patients admitted for GI complaints
		22 upper, 12 lower, and 13 upper and lower endoscopies performed, unclear distinction between primary and follow-up endoscopies
		Diagnostic yield 100%: Majority GVHD, gastritis, CMV, bacterial enteritis
Schwartz et al[18] (2001)	Prospective cohort 1985-1987 and 1996-1997	1102 patients with hematopoietic cell transplantation followed prospectively, of whom 75 developed severe GI bleed. Endoscopic evaluation included diagnostic and therapeutic procedures, however, number of procedures was unclear
		Diagnostic yield: Majority had multiple sites of bleed, caused by GVHD and peptic acid esophageal ulcers
Soylu et al[27] (2005)	Prospective cohort 1999-2005	451 patients with hematological malignancies, of which 32 developed overt GIB
		25 upper GI bleeding episodes, of which 8 EGDs were performed, remainder managed by supportive care. The other 7 patients had lower GI bleed episodes caused by neutropenic enterocolitis excluding the need for endoscopic procedures.
		Diagnostic yield 100% (8 endoscopies): Erosive gastritis (5/8), duodenal ulcers (3/8) in upper GI bleed

GI: Gastrointestinal; ERCP: Endoscopic retrograde cholangiopancreatography; EGD: Esophagogastroduodenoscopy; PEG: Percutaneous endoscopic gastrostomy; GVHD: Graft-vs-host disease.

Table 3 Thrombocytopenic precautions, therapeutic interventions, and bleeding adverse events

Study	Thrombocytopenic precautions	Therapeutic intervention	Bleeding Adverse events
	n = No. thrombocytopenic patients
Buderus et al[19]	Platelets < 30000/mm3: Biopsies not taken	4 PEG tube placements	None
	n = 12 (Platelets < 50000/mm3; 3 of 12 had platelets < 30000/mm3)	1 PEG tube removal
		2 sclerotherapies for varices
		6 NJ tubes placement
Chu et al[17]	Platelets < 20000/mm3: Biopsies not performed	None	None
	Platelet transfusion not a prerequisite, but made available
	n = 44 (Platelets < 40000/mm3; 25 of 44 had platelets < 20000/mm3)
Gorschlüter et al[20]	Platelets < 10000/mm3: Prophylactic platelet transfusion	8 endoscopic hemostasis in upper exam, including:	2 of 106 (1.9%) primary upper EGD had proven adverse events: hemorrhage induced by EGD (one stopped bleeding spontaneously and the other one required injection
	n = unknown	5 used fibrin glue
	Median platelets 23000/mm3	2 used fibrin glue plus epinephrine
		1 used epinephrine alone
		ERCP in 5 patients
		Duodenal tube placement in 8 patients	No ERCP-related adverse events
Kaur et al[22]	Platelets < 50000/mm3:	2 patients underwent successful electrocautery for bleeding ulcers	10 of the 31 patients in which endoscopies were performed had recurrent bleed at median of 7 d after index bleed (range 2-27 d), none readmitted
	Prophylactic platelet transfusion
	No target platelet count sought
	For all patients:
	Prophylaxis with H2 blockers or sucralfate or both		No adverse events as a result of endoscopy
	Hematopoietic cell progenitor support
	n = 27 (Platelets < 50000/mm3)
Kaur et al[23]	None	11 PEG tube placements	None reported
	n = unknown
Khan et al[24]	For platelets < 50000/mm3: Platelets transfused during procedure	None	GI bleeding adverse events occurred in 12 procedures out of 418 total procedures (2.9%). Thrombocytopenia was significantly associated (P < 0.01) with bleeding, occurring in 10 of the 12 procedures with bleeding adverse events
	n = 111 (Platelets < 50000/mm3)
			8 cases of bleeding events following EGD, of which there were:
			4 cases of duodenal hematomas that resolved with conservative management
			1 case requiring repeat endoscopy with electrocautery
			3 cases of acute GVHD managed conservatively
			4 cases of bleeding events following lower endoscopy"
			All due to acute GVHD
			Appear to have been managed conservatively
Park et al[21]	For platelets < 5000/mm3 or unstable (fever, hemorrhagic signs) patients with a platelet < 10000/mm3:	3 patients successfully treated with argon plasma coagulation for gastric angiodysplasia, hemoclips on colon ulcer, hemoclips on duodenal Dieulafoy’s lesion	1 death from massive GI bleed
			Re-bleed of Dieulafoy lesion, successfully treated by re-clipping
	Prophylactic platelet transfusion		No adverse events attributable to endoscopy
	n = unknown
Ross et al[25]	For platelets < 25-50000/mm3:	None	None reported
	Prophylactic platelet transfusion at discretion of endoscopist
	44 patients received prophylactic platelet transfusion
	n = at least 44 (Platelets < 25000-50000)
Schulenburg et al[26]	For platelets < 50000/mm3: Prophylactic platelet transfusion	None	None
	Platelet support to maintain count > 20000/mm3
	n = unknown
Schwartz et al[18]	For platelets < 50000/mm3:	2 attempted endoscopic hemostasis	No adverse events attributable to endoscopy reported
	No endoscopy if 50000/mm3 not reached	1 injection successful
	n = unknown	1 bipolar cautery plus injection that was unsuccessful and required surgery
Soylu et al[27]	For platelets < 20000/mm3:	None	No deaths or adverse events attributable to endoscopy
	Prophylactic platelet transfusion
	Active bleeding with higher platelet count also received prophylactic transfusion
	Severe thrombocytopenia (level not defined):
	EGD withheld in 17 of 25 upper GI bleeding episodes
	Colonoscopy withheld in 7 lower GI bleeding episodes
	n = unknown

GI: Gastrointestinal; EGD: Esophagogastroduodenoscopy; PEG: Percutaneous endoscopic gastrostomy; GVHD: Graft-vs-host disease.

Table 4 Neutropenic precautions and infectious adverse events

Study	Neutropenic precautions	Infectious adverse events
	n = No. of afebrile neutropenic patients
Buderus et al[19]	ANC < 1000/mm3 threshold:	One (2.1%) procedure-related adverse event:
	Upper endoscopies performed under “aseptic conditions” (not defined), appears that this did not include antibiotic prophylaxis	Fever and abdominal tenderness after colonoscopy
	Colonoscopies performed under antibiotic prophylaxis n = 10 (ANC < 1000/mm3)	Patient had not received antibiotic prophylaxis despite neutropenia (ANC 490/mm3); no explanation given in article
		Symptoms resolved in 2 d under IV antibiotics
Chu et al[17]	None	None
	n = unknown
Gorschlüter et al[20]	Neutropenia not defined	16 of 106 (15%) primary upper EGD: Fever within 48 h
	n = unknown	3 of 20 (15%) primary colonoscopies: Fever within 48 h
	Median WBC 1.5 G/l	Total # patients with fever following endoscopy: 19.
		5 of these died within 10 d.
		Not significantly different from # patients who died without having a fever following endoscopy.
		No ERCP-related adverse events
Kaur et al[22]	Neutropenia not defined	2 deaths due to sepsis
	n = unknown	No adverse events attributed to endoscopy
Kaur et al[23]	No neutropenic precautions taken	4 (36%) infectious adverse events total (both neutropenic and non-neutropenic)
	n = 4 (ANC < 1500/mm3)
		2 patients neutropenic at time of PEG placement.
		First patient had cellulitis and small abscess at PEG site, treated by removal of PEG
		Second patient had cellulitis at PEG site, treated by IV antibiotics
		2 patients non-neutropenic at time of PEG placement, but had neutropenia at the time of infection
Khan et al[24]	For ANC < 1000/mm3:	No infectious adverse events related to endoscopy.
	Broad-spectrum antibiotics prophylaxis	1 colonic perforation resulting in death
	n = 148 (WBC < 4000/mm3)
Park et al[21]	“Severe aplastic anemia” defined as bone marrow cellularity less than 25% and very low values for at least 2 of 3 hematopoietic lineages (including ANC < 500/mm3)	No adverse events attributable to endoscopy
	No precautions (no patients with fever)
	n = 28 (Severe aplastic anemia)
Ross et al[25]	None	None reported
	n = 0
Schulenburg et al[26]	Antibiotic prophylaxis during aplasia for all patients	None
	No extra prophylaxis for endoscopy
Schwartz et al[18]	None	No adverse events attributable to endoscopy
	n = unknown
Soylu et al[27]	Severe neutropenia (level not defined):	No adverse events attributable to endoscopy
	Withhold endoscopy in 17 upper and 7 lower GI bleed episodes
	n = unknown

EGD: Esophagogastroduodenoscopy; PEG: Percutaneous endoscopic gastrostomy; ERCP: Endoscopic retrograde cholangiopancreatography; GI: Gastrointestinal.