Q J Med 1996; 89:685-690
Push enteroscopy in the investigation of small-intestinal
disease
S. O ' M A H O N Y , A.J. MORRIS, M. STRAITON, L. MURRAY and J.F. MACKENZIE
From the Gastroenterology Unit, Glasgow Royal Infirmary, Glasgow, UK
Received 18 April 1996 and in revised form 14 June 1996
Summary
We report our experience with small-bowel push
enteroscopy in 50 patients. The indications for push
enteroscopy were: anaemia/occult gastrointestinal
bleeding (22 patients); overt gastrointestinal bleeding (17 patients); abnormal small-bowel radiology
(8 patients) and miscellaneous (3 patients). In those
with undiagnosed gastrointestinal bleeding/anaemia,
abnormalities were detected in 24/39 patients
(62%): small bowel arteriovenous malformations
(AVMs) were detected in 19 (49%), and five (13%)
had lesions in the upper gastrointestinal tract.
Seventeen patients had heater-probe ablation
therapy of vascular lesions: nine patients had smallintestinal lesions, four patients gastric lesions, and
four patients combined gastric and small-intestinal
lesions. In those with abnormal small-bowel radiology, abnormalities were detected in 6/8 patients.
We conclude that (i) push enteroscopy can establish
a diagnosis in a high proportion of patients with
gastrointestinal bleeding; (ii) heater-probe ablation
therapy of vascular lesions can be performed routinely at the time of enteroscopy; (iii) a significant
proportion of patients (9/50) referred for enteroscopy with undiagnosed gastrointestinal bleeding
have lesions in the stomach/proximal duodenum
missed at diagnostic endoscopy. Push enteroscopy
is a valuable diagnostic and therapeutic endoscopic
procedure.
Introduction
The small intestine is the part of the gastrointestinal
tract least accessible to endoscopic examination. The
proximal duodenum may be visualized by standard
oesophagogastroduodenoscopy (OGD), and approximately 30 cm of terminal ileum may be seen following intubation of the ileo-caecal valve following total
colonoscopy. The most important clinical indication
for visualization of the small intestine is obscure
gastrointestinal bleeding or iron-deficiency anaemia,
when a source of blood loss cannot be found in the
upper gastrointestinal tract or colon. It is now
recognized that small-intestinal lesions account for
a significant proportion of patients with obscure
gastrointestinal bleeding and iron-deficiency anaemia.1 For example, in a recent study of endoscopic
evaluation of patients with iron-deficiency anaemia,
standard OGD and colonoscopy failed to establish
a diagnosis in 38%; 2 as one third of these patients
had positive faecal occult blood tests, it must be
assumed that the small intestine is the source of
blood loss in these patients. Standard double-contrast
radiological examination of the small intestine, such
as enteroclysis, has a low yield of positive studies
(10%) in such patients;3 this is not surprising, as
vascular lesions, such as arteriovenous malformations
(AVMs), are the commonest source of small-intestinal
bleeding. 4
Operative enteroscopy can visualize the majority
of the small intestinal mucosa,5 but has the great
disadvantage of requiring a laparotomy. Sonde enteroscopy was first described in the 1970s,6 but its use
is still confined to a few specialist centres. Sonde
enteroscopy is a lengthy procedure, the instrument
tip cannot be controlled, and there is no facility for
biopsy or therapy. The sonde enteroscope is a long
(290 cm), narrow (5 mm) instrument which is passed
Address correspondence to Dr J.F. MacKenzie, Department of Medicine (Wards 8/9), Glasgow Royal Infirmary, 84 Castle
Street, Glasgow G4 OSF
© Oxford University Press 1996
686
S. O'Mahonyetal.
nasally.7 When the scope is in the stomach, it is
'piggy-backed' into the small intestine using a gastroscope.8 A balloon attached to the distal tip of the
instrument is inflated, and the scope is propelled
distally by peristalsis. The distal ileum is reached in
approximately 8 h, and the mucosa examined on
withdrawal of the instrument. The procedure makes
heavy demands on nursing time, and screening is
required. In experienced hands, the procedure has a
good diagnostic yield, 8 although its use is likely to
remain confined to a few specialist centres. 'Push'
enteroscopy was first described in 1983, using a
colonoscope to visualize the proximal jejunum. 9 The
adult colonoscope is too cumbersome and rigid for
this purpose, and the paediatric colonoscope, which
has a narrower, more flexible shaft, is preferred. The
paediatric colonoscope is short and 'floppy', allowing
for only limited intubation of the jejunum. For this
reason, dedicated push enteroscopes have been
developed, 10 with optimal characteristics of shaft
width, rigidity and optics for examination of the
small intestine. The Olympus Corporation have
developed two push enteroscopes, the Olympus
SIF-1O and SIF-1OL, with shaft lengths of 168 and
200 cm, respectively. The Pentax Corporation have
also developed a push enteroscope. These instruments have a standard biopsy/therapeutic channel
of 2.8 mm.
We have previously reported our initial experience
of small-bowel endoscopy using sonde enteroscopy,11'12 and we now describe our clinical experience with push enteroscopy over an 18-month
period.
caine throat spray. Most procedures were carried out
during routine endoscopy lists, or in the special
gastrointestinal investigation unit. Special preparation
was not required. Although X-ray screening is not
strictly necessary, we found it to be useful during
the learning phase of the procedure. Following
intubation of the pylorus and passage of the instrument to the distal duodenum, the enteroscope was
advanced through the duodeno-jejunal flexure and
thereafter advancement techniques were similar to
colonoscopy, the instrument progressing by judicious
rotation of the shaft, with shortening and straightening of the scope to facilitate advancement.
A duodenal-length overtube was used in a few
patients in an attempt to limit instrument looping in
the stomach. Small intestinal intubation length was
calculated by straightening the instrument to remove
the gastric loop, and subtracting 60 cm from the
depth inserted: 60 cm being the average distance
from incisors to pylorus with a straight endoscope.
The majority of procedures were performed by
a senior registrar (AJM) or consultant (JFMcK).
Procedure time in the majority of cases was less
than one hour.
Heater-probe therapy
The Keymed Unit with an 8F probe was used. The
instrument was set at a constant 10 J energy level
with a maximum of three applications per angiodysplasia allowed.
Results
Methods
Patients
Fifty patients underwent 61 push enteroscopy procedures over an 18-month period. Nine patients had
two procedures, and two patients had three procedures. The procedures were carried out from July
1993 to December 1994. Clinical indications for
push enteroscopy were as follows: iron-deficiency
anaemia/occult
gastrointestinal
bleeding
(22
patients); overt gastrointestinal bleeding (17 patients);
abnormal small-bowel radiology (8 patients) and
miscellaneous (3 patients). Patient details are shown
in Tables 1 to 3.
Enteroscopy technique
The Olympus SIF-10 enteroscope was used for all
examinations. This instrument has a shaft length of
168 cm, and a 2.8 mm biopsy channel. Enteroscopy
was performed under standard light intravenous
sedation with midazolam or diazepam, with ligno-
The median length of small intestine intubated was
60 cm (range 40-100 cm).
Anaemia/occult gastrointestinal blood loss
Results are shown in Table 2. Abnormalities were
found in 14/22 patients (64%). Seven patients had
small intestinal vascular lesions: five of these were
AVMs (Figure 1); one patient had hereditary haemorrhagic telangiectasia (HHT) with multiple smallintestinal vascular lesions, and a further patient had
'Blue-Rubber-Bleb' naevus syndrome, an inherited
condition characterized by cavernous haemangiomas
of the skin, gastrointestinal tract and other viscera.13
Two patients had NSAID-induced enteropathy, with
'red' spots and erosions in the small intestine.11
Three patients had vascular lesions in the stomach:
two had AVMs, and one had a 'watermelon' stomach
(multiple antral ectasias). One patient had a healed
gastric ulcer. Two patients had combined gastric/
small-intestinal AVMs. Enteroscopy was entirely
normal in seven patients.
687
Push enteroscopy
Table 1
Push enteroscopy in patients with obscure gastrointestinal blood loss/anaemia
No
Age/Sex
Enteroscopic findings
1
2
3*
4*
5
6
7*
8
9
10
11
47M
25F
65M
69F
26F
68F
66F
68M
81M
59F
51M
Healed GU
Jejunal erosions
3 jejunal AVMs
Normal
Normal
2 Gastric AVMs
Water-melon stomach
Normal
Duodenal AVMs
2 Gastric AVMs
Gastric AVMs, 1 duodenal
12
13
14*
15
16
17*
18
19
20
21
22*
46F
23F
Therapy
Yes
Yes
Yes
Partial
Yes
Yes
Outcome
Hb
Hb
Hb
Hb
No
Hb
Hb
Hb
Hb
Hb
Hb
stable
stable (misoprostil)
stable
stable
cause found
stable
stable
stable
stable
stable
stable; mitral valve replaced
AVM
71F
83F
36M
67M
63 F
63 F
52F
67F
58F
Normal
Blue rubber bleb syndrome
1 Gastric, 3 duodenal AVMs
Jejunal AVM
Normal
2 duodenal AVMs
2 duodenal AVMs
Normal
Jejunal telangiectasia
Normal
Jejunal erosions
Yes
Yes
Yes
Partial
Partial
Hb stable
Hb stable on oestrogen
Continued need for transfusion
Continued bleeding
No cause found
Not known
Not known
Not known
Not known
Not known
Hb stable (misoprostil)
M, male; F, female; GU, gastric ulcer; AVM, arteriovenous malformation; *patients who underwent two or more
procedures.
established in three: colonic carcinoma, Meckel's
diverticulum, and gastric ulcer (this last patient had
previously undergone both gastrocopy and push
enteroscopy).
Abnormal small-bowel barium radiology
See Table 3. Abnormalities were detected in 6/8
patients (75%): two patients had coeliac disease (one
had complicating ulcerative jejunitis), one patient
had probable jejunal lymphangiectasia, one patient
had extrinsic compression of the duodenum (which
was due to pancreatic pseudocyst), and in two
patients with proximal small-intestinal strictures, histology showed non-specific inflammatory changes.
Figure 1. Duodenal arteriovenous malformation (AVM).
Miscellaneous indications
Overt gastrointestinal bleeding
See Table 2. Abnormalities were found in 10/17
patients (59%). Six patients had small-intestinal vascular lesions: five had AVMs, and one patient had
hereditary haemorrhagic telangiectasia. Two patients
had lesions in the stomach/duodenal bulb: one
patient had a gastric AVM, and one patient had
duodenitis. Two patients had combined gastric/small
intestinal AVMs. In the seven patients with normal
enteroscopy, a definitive diagnosis was subsequently
One patient had a protein-losing enteropathy of
unknown cause; one patient had rheumatoid arthritis
and a high faecal fat excretion, and one patient had
abdominal pain with dilated loops of small intestine
on plain abdominal X-ray. Enteroscopy in all three
was normal.
Heater-probe ablation of vascular lesions
We have described our experience with this therapy
elsewhere,14 including many of the patients in this
5. O'Mahony etal.
688
Table 2
Push enteroscopy in patients with overt gastrointestinal bleeding
Age/Sex
Enteroscopic findings
Therapy
Outcome
1
2
71M
78M
Normal
Normal
Jejunal AVM
Jejunal AVMs
Failed
Yes
33F
67F
Jejunal telangiectasia
Normal
Failed
8*
43M
Yes
9
32M
Gastric and small intestinal
AVMs
Duodenitis/healed DU
Further bleeding
Found to have colonic
carcinoma
Found to have Meckel's
diverticulum
Continued bleeding
Hb stable; no further
bleeding
Continued bleeding
Gastric ulcer found when
patient rebled
Still transfusion-dependent
3
18M
Normal
4*
5*
80F
63M
6
7
10
32M
Gastric AVM
Yes
11*
67M
Yes
12
13
70F
59F
4 gastric/1 duodenal/1
jejunal AVMs
Normal
4 jejunal AVMs
14
15*
51F
74M
2 jejunal AVMs
2 jejunal AVMs
Failed
One AVM ablated; failure of
ablation in second AVM
16
17
72 F
Normal
Normal
No
74F
Yes
Hb stable; no further
bleeding
Hb stable; no further
bleeding
Further bleeding;
commenced on oestrogens
Not known
Hb stable; no further
bleeding
Not known
Not known
Not known
Not known
M, male; F, female; DU, duodenal ulcer; Hb, haemoglobin; *patients who underwent two or more procedures.
series. Seventeen patients in this series underwent
heater-probe ablation therapy of vascular lesions.
Discussion
We established a diagnosis in 31 of our 50 patients
(62%). It should be pointed out, however, that in 9
of these 31 patients, the lesions discovered could
have been identified by standard gastroscopy. Other
investigators have reported a similar incidence of
lesions missed by gastroscopy.15'16 Some authors
have therefore argued that repeat gastroscopy should
be performed in patients with obscure gastrointestinal
bleeding. 17 Push enteroscopy therefore identified
small-intestinal abnormalities in 22/50 patients
(44%). As might be expected, the diagnostic yield
was particularly high in patients with abnormal
small-intestinal radiology (with the abnormalities in
the proximal small intestine). In the patients with
overt or occult gastrointestinal bleeding, a small
intestinal source was identified in 16/39 patients
(41%), with AVMs being the most common lesion.
Other investigators have reported a similarly high
frequency, of AVMs detected by push enteroscopy in
patients with obscure gastrointestinal bleeding/
anaemia.16'17
We have previously reported on our experience
with sonde enteroscopy;11'12 although valuable in
the investigation of patients with suspected smallintestinal bleeding, the technique is limited by the
inability to control the scope tip deflection or to
carry out mucosal biopsy or therapy. Nevertheless,
sonde enteroscopy gives an additional 26% yield
over push enteroscopy,18 and therefore it is used in
a few specialized centres. Push enteroscopy, however, is relatively quick and straightforward. Most
procedures can be performed during the course of a
routine endoscopy list, and the procedure is generally
we 11-tolerated. Another advantage of push enteroscopy compared to sonde enteroscopy is the ability
to carry out mucosal biopsy and perform therapeutic
manoeuvres. Apart from ablation of AVMs, other
therapeutic applications of push enteroscopy include
percutaneous endoscopic jejunostomy, 19 stricture
dilatation, polypectomy 15 and ERCP following Rouxen-Y reconstruction.20 The chief limitation of push
enteroscopy is the fact that the limit of examination
is generally the mid-jejunum. Depth of intubation
may be increased by the use of an overtube. The
689
Push enteroscopy
Table 3
Push enteroscopy in patients with abnormal small bowel barium radiology
No
Age/Sex
Clinical problem/
Radiology
Enteroscopic findings
Biopsy
Outcome
1
65M
Well on gluten-free diet
28M
Multiple jejunal
strictures/ulceration
Normal
Ulcerative jejunitis
2
Abdominal pain/jejunal
ulceration
Abdominal pain/
thickened small bowel
loops
Normal
3
21M
Jejunal ulceration
Inconclusive:
?lymphangiectasia
4
25F
Protein-losing
enteropathy/diffusely
abnormal mucosa
Diarrhoea/thickened
duodenal folds
Normal
Normal
5
29F
51F
Extrinsic duodenal
compression
Multiple ulcers first and
second part of
duodenum
Normal
6
7
67M
8
62 M
Vomiting/duodenal
stricture
Recurrent duodenal
ulceration/multiple
ulcers into distal
duodenum
Vomiting/stricture
duodenojejunal
flexture
Diarrhoea/thickened
small bowel loops
Radiological
abnormalities
confirmed to be
artefactual
Continued
gastrointestinal protein
loss
Radiological
abnormalities
confirmed to be
artefactual
Compression due to
pancreatic pseudocyst
Not known
Chronic inflammation/
ulceration ?Crohn's
Proximal jejunal
stricture
Chronic inflammation
?Crohn's
Not known
Jejunal ulceration/
stricture
Subtotal villous atrophy
Well on gluten-free
diet + steroids
instrument which we used has a shaft length of
1 68 cm; an instrument with a shaft length of 200 cm
is now available (the Olympus SIF-10L). It is likely
that with increased experience of push enteroscopy,
particular manoeuvres and 'tricks of the trade' (e.g.
changes in patient positioning, rotation/shortening of
the scope) will become established as means of
achieving optimal small-intestinal intubation. Longer
enteroscopes are being developed, and preliminary
studies suggest that intubation to the distal jejunum
is possible with these instruments.21 We were able
to apply heater-probe ablation therapy successfully
in 17 patients, although ablation was incomplete in
four of these patients. Ablation therapy could not be
carried out in three patients. We have reported on
enteroscopic heater-probe therapy of AVMs.14
The patients described in this series came from
both our own unit and other hospitals, and therefore
had varying degrees of investigation prior to enteroscopy. Differences in the extent of pre-enteroscopy
investigation probably account for the wide variation
in diagnostic rates reported in the literature for both
push and sonde enteroscopy. In a recent American
study, 100 consecutive patients with iron-deficiency
anaemia were investigated by endoscopy (gastroscopy and colonoscopy), and lesions potentially
responsible for blood loss were detected in 62%. 2 In
this series, small-bowel barium radiology was per-
formed in 26 of the 38 patients with negative
endoscopic studies, and was normal in all. A proportion of these patients undoubtedly had smallintestinal blood loss, as faecal occult blood tests
were positive, and gastroscopy and colonoscopy
were negative. (It should be noted, however, that in
this series, duodenal biopsies were not routinely
performed, so it is possible that some of these
patients had coeliac disease). How should the clinician proceed in cases of iron-deficiency anaemia
and gastrointestinal haemorrhage of obscure origin?
If gastroscopy (with duodenal biopsy) and colonoscopy are negative, then push enteroscopy is the
next reasonable step. Should all of these investigations fail to reveal a source of blood loss, then
sonde enteroscopy is indicated in patients with slow
blood loss, and angiography or operative enteroscopy
in patients with life-threatening bleeding.
Sonde enteroscopy is a major undertaking for both
endoscopist and patient. There is a steep learning
curve,22 and it would therefore seem sensible for this
procedure to be available on a regional basis, with
experience concentrated in a few centres to maintain
expertise. Push enteroscopy is a procedure which
can be quickly learned by the experienced endoscopist, and which can be carried out during the course
of routine endoscopy lists. We believe that push
enteroscopy should be used early in the evaluation
690
S. O'Mahonyet al.
of patients with suspected small-bowel blood loss or
abnormal small-bowel radiology.
10. Shimuzu S, Tada M, Kawai K. Development of a new
insertion technique in push-type enteroscopy. Am
J Gastroenterol1987; 82:844-7.
Acknowledgments
11. Morris AJ, Madhok R, Sturrock RD, Capell HA, MacKenzie
JF. Enteroscopic diagnosis of small bowel ulceration in
patients receiving non-steroidal antiinflammatory drugs.
Lancet!991; 337:520.
JFMcK was in receipt of a grant to develop a novel
clinical service from the Scottish Home and Health
Department. SO'M was seconded to Glasgow Royal
Infirmary Castroenterology Unit for a two-month
sabbatical period from Northern & Yorkshire Region.
References
1. Thompson JN, Hemingway AP, McPherson CAD, Rees HC,
Allison DJ, Spencer J. Obscure gastrointestinal haemorrhage
of small-bowel origin. BrMedJ 1984; 288:1663-5.
2. Rockey DC, Cello JP. Evaluation of the gastrointestinal tract
in patients with iron-deficiency anaemia. N Engl J Med
1993; 329:1691-5.
3. Rex DK, LappasJC, Maglinte DDT, Malczewski MC,
Kopecky KA, Cockerill EM. Enteroclysis in the evaluation of
suspected small intestinal bleeding. Gastroenterology 1989;
97:58-60.
4. Foutch PC. Angiodysplasia of the gastrointestinal tract. Am
J Gastroenterol 1993; 55:807-18.
5. Szold A, Katz LB, Lewis BS. Surgical approach to occult
gastrointestinal bleeding. AmJ Surg 1992; 163:90-3.
6. Tada M, Akasaka Y, Misaki F, Kawai K. Clinical evaluation
of a sonde-type intestinal fiberscope. EnGtoscopy 1977;
8:33-8.
12. Morris AJ, Wasson LA, MacKenzie JF. Small bowelenteroscopy in undiagnosed gastrointestinal blood loss. Gut
1992; 33:887-9.
13. Berlyne GM, Berlyne N. Anaemia due to 'Blue-Rubber-Bleb'
naevus disease. Lancet^960; ii:1275-7.
14. Morris AJ, Mokhashi M, Straiton M, Murray L, MacKenzie
JF. Push enteroscopy and heater probe therapy for small
intestinal bleeding. Gastrointest Endosc (in press).
15. Davies GR, Benson MJ, Gertner DJ, Van Someren RMN,
Rampton DS, Swain CP. Diagnostic and therapeutic pushtype enteroscopy in clinical use. Gut 1995; 37:346—52.
16. ChongJ, Tagle M, Barkin JS, Reiner DK. Small bowel pushtype fiberoptic enteroscopy for patients with occult
gastrointestinal bleeding or suspected small bowel
pathology. Am J Gastroenterol 1994; 89:2143-6.
17. Foutch PG, Sawyer R, Sanowski RA. Push-enteroscopy for
diagnosis of patients with gastrointestinal bleeding of
obscure origin. Gastrointest Endosc 1990; 36:337-41.
18. Bemer JS, Mauer K, Lewis BS. Push and sonde enteroscopy
for the diagnosis of obscure gastrointestinal bleeding. Am
J Gastroenterol! 994; 89:2139-42.
19. Shike M, Wallach C, Likier H. Direct percutaneous
endoscopic jejunostomies. Gastrointest Endosc 1991;
37:62-5.
7. Tada M, Shimuzu S, Kawai K. A new transnasal sonde type
fiberscope (SSIF type VII) as a pan-enteroscope. Endoscopy
1986; 18:121-4.
20. Gostout CJ, Bender CE. Cholangiopancreatography,
sphincterotomy, and common duct stone removal via Rouxen-Y limb enteroscopy. Gastroenterology 1988; 95:156-63.
8. Lewis BS, Waye JD. Chronic gastrointestinal bleeding of
obscure origin: role of small bowel enteroscopy.
Gastroenterology 1988; 94:1117-20.
21. Barkin JS, Lewis BS, Reiner DK, Waye JD, Goldberg Rl,
Phillips RS. Diagnostic and therapeutic jejunoscopy with a
new, longer enteroscope. Gastrointest Endosc 1992;
38:55-8.
9. Parker HW, Agayoff JD. Enteroscopy and small bowel
biopsy utilizing a peroral colonoscope. Gastrointes Endosc
1983; 29:139-40.
22. Lewis BS, WayeJD. Small bowel enteroscopy in 1988: Pros
and Cons. AmJ Gastroenterol 1988; 83:799-802.