Colonic Vascular Ectasias and Aortic Stenosis: Coincidence or Causal Relationship?
Robert J. Greenstein, MD, New York, New York
A. James McElhinney, MD, Bronx, New York Devaprasad Reuben, RRA, New York, New York Adrian J. Greenstein, MD, New York, New York
Controversy exists as to whether the relationship between aortic stenosis and colonic vascular ectasias is causative or coincidental [I]. We reviewed 11 patients with a diagnosis of vascular ectasia and gastrointestinal hemorrhage and evaluated their cardiologic diagnosis. We studied the pulse wave pattern delivered to the cecum in two patients with ectasias and three control subjects. In a separate study, we compared the incidence of gastrointestinal hemorrhage in 3,623 patients with aortic or mitral stenosis admitted to the Mount Sinai Hospital over a 10 year period. This statistical association of gastrointestinal hemorrhage with aortic stenosis has been previously noted [2]. In our patients, gastrointestinal bleeding was significantly more common in those with aortic stenosis (21 of 1,811 patients) than in those with mitral stenosis (1 patient) (chi-square = 18, p
From the Departments of Surgery of The Mount Sinai Medical Center, New York, New York, and the Veterans Administration Medical Center, Bronx, New York. Requests for reprints should be addressed to Robert J. Greenstein. MO, S.A. Berson Research Laboratory, Veterans Administration Medical Center, 130 West Kingsbridge Road, Bronx, New York 10466.
Volume 161, March 1966
trointestinal diagnostic studies, and pathologic studies of the resected specimens. In three resected right hemicolectomy specimens, injection studies were performed by Dr. R. Sammartano courtesy of Dr. S. J. Boley of Montefiore Medical Center, New York, according to their published methods [I]. The diagnosis of aortic stenosis was based on previously described criteria [2]. These include combinations of a systolic ejection murmur (greater than or equal to III/VI) heard at the base and not corrected by blood replacement, calcification of the aortic leaflets, ancillary evidence on electrocardiogram, decreased or absent As. carotid pulsus parvus et tardus, and left ventricular hypertrophy on chest radiograph without hypertension, Arterial pulse waves were recorded at the time of surgical excision of the right side of the colon in two patients with a presumptive diagnosis of vascular ectasias. The control subjects consisted of three persons undergoing right hemicolectomy, two for carcinoma and one for Crohn’s disease. In each patient, both a systemic artery and the ileocolic artery were cannulated. Recordings of pulse waves in both arteries were made using a Statham P23 DB@ pressure transducer and recorded onto E for M Simultracee recording paper. Study 2: The yearly incidence of patients discharged from the Mount Sinai Hospital with the diagnosis of aortic or mitral stenosis during the period 1969 through 1978 was obtained from the computer files of the hospital. The records of 28 patients with associated gastrointestinal hemorrhage were retrieved after cross-indexing. The microfilms of these 28 patients were reviewed for age, sex, drug history, and cardiologic diagnosis. Investigations and surgical procedures performed for gastrointestinal hemorrhage were noted. The chi-square test was used for statistical analysis [3].
Results Study 1: Eleven patients had the diagnosis of vascular ectasia established at the Mount Sinai Medical Center between 1970 and 1978. Six patients
347
Greenstein et al
78 y o 9
Vascular
Ectasias
!
Ilio-colic Artery
Radial Artery a
t
Figure 1. Diagnostic criteria for vascular ectasias of the colon in 17 patients. The most useful diagnostic modality is angiography. Colonoscopy hadless value. The sensitivity of the latex injection technique of Boiey et al [ l] is obviously greater than routine formaldehyde fixation. Asterisk indicates two patients wlth bleeding.
were seen at the Mount Sinai Hospital, four at the Bronx Veterans Administration Medical Center, and a single patient at Elmhurst City Hospital. The mean age of these patients was 71 years (range 53 to 84 years). The youngest was on long-term hemodialysis. The mean duration of bleeding was 9 months. The mean number of blood units transfused was 32 (range 5 to 116 units) and the mean admission hemoglobin, 6.9 g. The diagnosis of aortic stenosis was established in eight patients by previously stated criteria [2]. The remaining three had limited documentation, but all three had a systolic ejection murmur at the left sternal border or aortic area radiating to the neck or the carotid vessels. No patient had left ventricular catheterization, aortic valve operation, or an autopsy. The preoperative diagnosis of vascular ectasia was made in 10 of the 11 patients by angiography. In the other patient, it was made by visualizing an actively bleeding ectasia during colonoscopy (Figure 1). All 11 patients underwent laparotomy and right hemicolectomy. Three specimens had latex fixation before histopathologic examination [I]. In all three, vascular ectasias were demonstrated. The remaining eight had routine formaldehyde fixation before histopathologic examination. Ectasias were demonstrated in only two of these eight patients. After right hemicolectomy, three patients rebled. One had a further resection (subtotal colectomy with ileoproctostomy) and the others were managed conservatively with oral iron and transfusion.
348
‘3
/
i
54 v.0. P
Vascular
l---l
Ectasias
set---i
Figure 2. The configuration of these two pulse waves (top and bottom) show loss of anacrotic and dicrotic notches. However, pulsus parvus et tardus Is not well demonstrated. These readings suggest nonhemodynamicalfy significant aortic stenosis.
The pulse wave pattern of two of the patients with ectasias are shown in Figure 2. The pathognomonic loss of both the anacrotic and dicrotic notches is noteworthy. In contrast, patients with cancer and Crohn’s disease show normal pulse waves (Figure 3). Both the anacrotic and dicrotic notches are preserved, even in the mobilized distal ileocolic artery. Study 2: The total number of patients discharged from the Mount Sinai Hospital with the diagnosis of aortic stenosis and mitral stenosis during the decade from 1969 through 1978 was obtained from the computer files of the hospital. There were 1,811 with
The American Journal of Surgery
Colonic EctasiasandAorticStenosis
72
Ca Cecum
y o. d”
t
I
-~ve~~;-qyJy~ ’ Ilio-colic
/\
Artery
/ ‘,
set----I
+I
53 y.o
63 mm Ha
y o. d”
I
I
Ilio-colic Artery
Awta
8
+
Cecal
I
I I
1 I
I
/
I set +
Crohn’s
lleitis
Radial
Co
I
c?
+
I sech
Figure 3. Despite patient age and locatkm, theee pulse waves are within normal Ilmits, demonstrating that there Is no maJor change In conflguratlon or pulse wave pattern In the terminal superior mesenterlc artery. Ca = cancer.
aortic stenosis and coincidentally, an almost identical number (1,812 patients) with mitral stenosis (Table I). Twenty-eight of 3,623 patient charta that showed both valvular disease and gastrointestinal hemorrhage were analyzed. There were 25 patients with aortic stenosis and 3 with mitral stenosis. The following six patients, two with disease of the mitral valve and four with aortic’ stenosis, were excluded from the final analysis. With mitral valvular disease, one patient had both mitral stenosis and incompetence. She was a 45 year old woman receiving prophylactic warfarin (Coumadina) anticoagulation therapy. She had an endoscopic sigmoid poly-
Vqlume 151, Mwch 1986
pectomy and stopped bleeding when her coagulation parameters were corrected. The second was a 70 year old woman with an endoscopic diagnosis of acute erosive gastritis. She had been taking’large doses of aspirin for an acute herpetic radiculopathy. Of the four patients with aortic stenosis, one was a 25 year old Hispanic man with well-documented Rendu-Osler-Weber syndrome in addition to his aortic stenosis and gastrointestinal hemorrhage. The additional three patients with aortic stenosis who were excluded, included two with colonic cancer and one with a jejunoileai arteriovenous malformation. In the remaining 22 patients, 21 with aortic stenosis and 1 with mitral stenosis, the discharge
349
Greenstein et al
TABLE I
Numbsr of Pstlents in Each Year (1969 to 1978) Year
Aortic Stenosis
TABLE 11 Dlagnosls at Otscherge In 21 PUients With Aortic Stenosis and Gastrointestinal Hemorrhage
Mitral Stenosis
Patients Diagnosis
1969 1970 1971 1972 1973 1974 1975 1976 1977 1978
127 162 208 166 171 183 199 193 231 171
133 147 197 158 118 182 247 204 221 205
Total
1.811
1.812
diagnosis did not adequately explain their gastrointestinal hemorrhage (Table II). The mean age of the 21 patients with gastrointestinal hemorrhage and aortic stenosis was 74 years (range 58 to 86 years). There were 13 men and 11 women. Before 1972, four patients had one or more gastric operations for gastrointestinal hemorrhage (Table III). A comparison of the 21 of 1,811 subjects with gastrointestinal hemorrhage and aortic stenosis with the 1 of 1,812 subjects with mitral stenosis using the chi-square test showed a difference that was highly statistically significant (chi-square = 18 with 1 degree of freedom, p
TABLE III
Upper gut Gastric telangectasias Antrai erosion Hiatus hernia Lower gut Cecal ectasias (by colonoscopy) Diverticulosis Cryptogenic
4 1 2
No No No
2 2 10
Yes No Yes
Total
21
. *
replacement might be an indicated procedure of choice for colonic bleeding from vascular ectasias [IO]. Speculation as to whether this relationship is ~oincident~ or causative has continued [1,6]. Our studies in large part confirm previous ones in terms of the age of onset of ectasias, presentation, difficulties with diagnosis, and difficulties with routine histopathologic examination. The great value of angiography and latex injection to ascertain the pathologic diagnosis are confirmed. The latter must be considered the gold standard for the diagnosis of this entity. It is the role that aortic stenosis plays that is most controversial. Although the association is well established, the possible connection remains speculative. Boley et al [If have stated that both aortic stenosis and ectasias occur commonly in the elderly, but they believe this is not a causative relationship; instead, they suggest that the connection might be aortic valvular-induced cardiovascular compromise. Bleeding occurs because of ischemia in the monoendothelial layer separating the vascular from the colonic lumen. It seems improbable to us that a vascular endothelial cell could develop ischemia, and even less so when there is increased arterial
Result
Patient
Oaeration
Date
1
Vagotomy and pyloropiasty Subtotal gastrectomy with Billroth II gastrectomy Billroth II gastrectomy
1967 1960
Ongoing bleeding Ongoing bleeding
1971
Vagotomy and pyloroplasty Billroth II gastrectomy Vagotomy and Billroth II gastrectomy
1971 1972 1972
Persistent anemia with guiac posit&e stool Still bleeding Still bleeding Persistent black stool & anemia
3 4
350
Active Bleeding
Results of Gastric Operations In Four Patients Wlth Aortlc Stenosis and Recurrent Gastrointestinal Hemorrhage”
2
l
(n)
All gastric operations performed before 1972.
The American Journal of Surgery
Colonic Ectasias and Aortic Stenosis
blood coursing through, as occurs with an ectatic vessel. We believe that this relationship is causative. The studies we have performed show that the abnormal pulse wave that is associated with aortic stenosis is found in the terminal branch of the ileocolic artery. This differs from the systemic configuration that is consistently present in the ileocolic artery in the absence of aortic valvular disease. It is how this abnormality might cause the ectasias that remains speculative. Chronic intermittent venous obstruction is offered as the most plausible explanation for the development of ectasias by Boley et al [I]. They cite Laplace’s law for increased tension in the cecum (T = ?rDP, where T = tension, D = diameter, and P = intraperitoneal pressure) as corroborating their hypothesis. Yet even a superficial examination of the entire bowel shows that the cecum has, by far, the thinnest wall and most attenuated musculature. One would expect that if intermittent muscular venous obstruction was responsible for the development of ectasias, they should occur most commonly in areas of the gut where the musculature was thickest, notably the sigmoid colon or stomach, not thinnest, as in the cecum. We suggest that it is the thinness of the wall of the cecum that is responsible for the development of ectasias. The lack of extrinsic support of the vascular architecture allows the abnormal pulsus parvus et tardus pulse wave striking the undersupported vasculature of the cecum to gradually distend the submucosal venous plexus. Thus, we suggest the cause of ectasias should be attributed to the pathologic abnormalities in the inflow pulse wave pattern rather than intermittent obstruction in the outflow tract. However, this speculation is very difficult to confirm. The fact that the ectasias classically occur in the elderly implies that many years of abnormal wave patterns striking the submucosal venous plexus in the thinner wall of the cecum are required for ectasias to develop (the number of heart beats that occur in a 71 year old is approximately 3 X log). Therefore, even if an animal model with a large cecum and nonhemodynamically significant aortic stenosis could be obtained, the possibility of confirming this hypothesis would remain remote at best.
Volume 151, March 1986
Summary Eleven patients with vascular ectasias of the colon and associated gastrointestinal hemorrhage were evaluated. All had the clinical features associated with aortic stenosis. In two patients, the configuration of the pulse wave in the mesenteric vessel was studied. In both, the abnormal peripheral pulse wave pattern associated with aortic stenosis was also transmitted to the ileocolic artery, where it differed quite clearly from the pattern in control patients. In a parallel study, the computer records of 3,623 patients with aortic or mitral stenosis admitted to the Mount Sinai Hospital over a 10 year period were reviewed for the presence of cryptogenic gastrointestinal hemorrhage. Twenty-one of 1,811 patients with aortic stenosis but only 1 of 1,812 patients with mitral stenosis had concomitant gastrointestinal hemorrhage (chi-square = 18, p
351