- Email: [email protected]
Barium enema or computed tomography for the frail elderly patient?
Clinical Radiology (1993) 48, 48 51
Barium Enema or Computed Tomography for the Frail Elderly Patient? J. J. DAY*, A. H. F R E E M A N , N. K. CONI* and A. K. D I X O N
Departments of Radiology and *Geriatric Medicine, Addenbrooke's Hospital and the University of Cambridge, Cambridge In order to determine whether abdominopelvic computed tomography (CT) offers an alternative to barium enema (BE) for the investigation of the large bowel in frail elderly patients, we have assessed and compared the results of both tests performed in each patient. Thirty-seven patients aged 71 to 88 (mean 80) with a history or clinical findings suggestive of large bowel disease were studied. The seven colonic neoplasms in this group were demonstrated by both techniques (apart from one patient who did not undergo BE as CT had shown an obstructing lesion). CT showed useful additional data in four patients (e.g. abdominal wall involvement) and demonstrated numerous extracolonic lesions (e.g. ovarian carcinoma). However, CT did raise the possibility of a large bowel neoplasm in four patients where none was shown by BE and missed one case of Crohn's disease. CT was the preferred test amongst the 25 patients where the acceptability of the two techniques could be compared. Only in 16 patients were the BE studies adjudged to be of good quality. CT should be the initial investigation of the large bowel in frail elderly patients requiring inpatient bowel preparation; the more unpleasant BE could be reserved for those cases where ~ T is equivocal or severe symptoms are unexplained. Day, J.J., Freeman, A.H., Coni, N.K. & Dixon, A.K. (1993). Clinical Radiology 48, 48-51. Barium Enema or Computed Tomography for the Frail Elderly Patient?
Accepted for Publication 21 January 1993
Barium enema (BE) is recognized as the standard radiological investigation for suspected large bowel disease. However, it is often a difficult procedure in elderly patients and each referral should be carefully considered [1]. Bowel preparation is frequently unsatisfactory and distressing. So much so that it is our usual practice (in common with many other geriatric units) to admit most of these patients for the preparation and the enema itself [2]. This adds to the expense of the investigation. Incontinence [3] and poor mobility may hinder the procedure. Furthermore, about one-third of BEs in this age group may be inconclusive due to the presence of faeces or incomplete demonstration of the whole of the large bowel [4]. Computed tomography (CT) has been shown to be capable of identifying most carcinomas of the large bowel [5,6], especially when these are large. CT can also demonstrate a wide range of mucosal abnormalities, such as diverticular and inflammatory bowel disease [7,8]. CT also provides additional information about other intra-abdominal structures which might be of value in these elderly patients, w h o often present with rather nonspecific symptoms and signs [9]. In order to see whether CT (a less invasive test) could replace BE in patients with features suggestive of large bowel disease, we performed CT on most patients who were referred for BE from the geriatric unit over a 5 month period. P A T I E N T S AND M E T H O D S Thirty-seven patients who were referred for BE from the Care of the Elderly Department at Addenbrooke's Correspondence to: Dr A. K. Dixon, University Department of Radiology, Addenbrooke's Hospital, Cambridge CB2 2QQ.
Hospital over a 5 month period were entered into the study. After obtaining verbal consent for an additional test to be done, CT was booked to coincide with the day of admission for the start of the bowel preparation (48 h before BE). On arrival the patients were given regular drinks of dilute oral contrast medium along with the laxative. Later that day, immediately before CT, they were given yet more oral contrast medium, the aim being to opacify the whole gastrointestinal tract. CT was performed using 1 cm thick slices at 2 cm increments through the whole abdomen and pelvis. In our unit, the decision as to whether to give intravenous contrast medium or to obtain further views (intervening slices, thin slices, decubitus images, delayed images, etc.) is up to the radiologist performing the CT examination; such additional procedures were performed as deemed appropriate. It is worth emphasizing that no patient received rectal contrast medium before CT. The CT study was reported, with special attention to the colon, by the registrar, senior registrar or consultant in charge of the session; registrars have to get their work checked by a more senior colleague. However, the CT report was withheld until the BE had been done. BE was carried out 2 days later, after 48 h of bowel preparation. A conventional double contrast technique was attempted wherever possible. The various manoeuvres available to facilitate complete opacification of the large bowel were used as appropriate. Again the procedure was performed by a registrar, senior registrar or a consultant; a registrar's report had to be checked by a more senior colleague. The report was made without access to the CT findings. The BE images were retrospectively reviewed and graded (AF) for quality into nondiagnostic, poor (allowing only a qualified report) or
BARIUM ENEMA OR CT IN THE ELDERLY.9
Fig. 1 - A 73-year-old man with anaemia, altered bowel habit and weight loss. CT shows clear evidence of small bowel dilatation culminating in a caecal mass. Note the thickened bowel wall around an irregular central area of contrast medium at the expected site of the caecum. This lesion was successfully removed at surgery. Barium enema was not performed.
good, according to the adequacy of the preparation, the degree of colonic visualization and the extent of barium retention. Following BE the results of both tests were made available and the management of the patients proceeded along usual lines. Only at this stage were the radiologists invited to comment on the two tests together (in cases of discrepancy). The patients went home as soon as possible after BE, with outpatient follow-up if necessary. Some patients remained in hospital for further treatment. The patients were asked by the referring clinician (within 3 weeks of the procedures) about their perception (pleasant/unpleasant) of the two techniques; these findings were analysed by the Chi-squared test. They were also asked to score each test on a 10 cm visual scale (where zero represented the worst experience ever and 10 the best); these scores were compared by the Wilcoxon's test for pair differences. At the end o f the study we analysed the radiological reports in each patient, along with the subsequent clinical course. Follow-up was arranged at 3 months, either through outpatients or the general practitioner so that any change in clinical condition could be assessed. In particular we classified the colonic findings into likely neoplasm, possible neoplasm, diverticular disease, inflammatory bowel disease and normal. RESULTS The age of the 37 patients ranged from 71 to 88 (mean 80) years. CT was possible in 36 o f the 37 patients; claustrophobia was the reason for the single failure. BE was technically impossible and completely abandoned in three patients (one at the second attempt). In four other patients the first BE attempt proved unsuccessful; images were only obtained at a second attempt. BE was not performed in one further patient as CT (Fig. 1) had shown a large caecal lesion which was causing small bowel obstruction. Amongst the 33 BEs which were available for review, two were deemed to be of non-diagnostic quality, 15 were judged to be of poor quality (although a qualified
49
Fig. 2 - An 86-year-old lady with constipation, abdominal pain and weight loss. CT shows a 2.4 cm mass (arrows) of apparent soft tissue in the caecum which also contains air and faeces. This was reported as a possible tumour. In retrospect it has the typical appearances of the normal ileo-caecal valve. Better opacification with oral contrast medium might have prevented this error of interpretation. The subsequent barium enema was normal with a normal ileo-caecal valve at this site.
conclusion could be made) and 16 were considered of good quality. Comparison of the acceptability of the two techniques was not possible in every patient because of dementia, illness, etc. Twenty-five patients were able to provide a simple comparison: 14 considered BE unpleasant (three said they would never consent to it being done again); CT was considered unpleasant by three (Z:= 10.75, P < 0.005). Twenty-one patients could give visual scores; these were higher for CT (mean 5.19 cm) than BE (mean 3.90 cm); 11 patients scored CT higher; eight patients gave both techniques the same score; two patients gave BE higher scores. The difference in these scores reached statistical significance (P < 0.05) on testing for pair differences. After 3 months follow-up, seven colonic (caecum 2, ascending colon 2, descending colon 1, rectosigmoid 2) neoplasms had been discovered amongst the 37 patients. All seven had been reported as a likely malignancy in the CT report; in one of these (Fig. 1), the presence of severe small bowel obstruction led to surgery without BE being performed. BE demonstrated the other six neoplasms, although in one patient the first BE had to be abandoned and the lesion in the ascending colon was only demonstrated on a second BE (in the knowledge of the positive CT findings). CT raised the possibility of a neoplasm in four further patients where no lesion was identified at BE. In one of these the radiologist over-reported the normal ileo-caecal valve (Fig. 2); in another there was concern about a possible lipoma of the ileo-caecal valve. In a third, the tissue planes between the rectum and vagina were indistinct; it transpired that the patient had previously received a caesium implant for carcinoma of the cervix. In the fourth patient a tumour could not be excluded in and amongst a segment of sigmoid colon affected by diverticular disease. There were no such false positive BE reports. Diverticular disease was mentioned in 19 of the 35 BE reports. On retrospective review of the BE and CT films there was quite close correlation in the recognition of diverticular disease between the two techniques. There were three patients where isolated areas of diverticular disease were only shown by BE. There was one patient,
50
CLINICAL RADIOLOGY
(a)
adjacent abnormal fat was visible on the CT images. In a patient with benign haemangiomatosis o f the rectum shown by CT the first BE was unsuccessful. The abnormality was only recognized in retrospect on the second BE. CT identified numerous extra colonic lesions. In two of the seven patients with colonic neoplasms CT suggested the presence o f hepatic metastases; in another, with a sigmoid carcinoma, CT showed ascites and an obstructed left ureter; in a fourth, CT showed abdominal wall involvement (Fig. 3). Amongst the other 30 patients CT revealed a large cystic ovarian mass in one patient in whom BE was impossible. In another, a complex pelvic paracolic mass, with air, fluid and calcium within, was demonstrated by CT; an infected gynaecological lesion was suggested; BE suggested a diverticular abscess or Crohn's disease; at surgery an infected ovarian dermoid was removed. In a further patient a possible mass in the right lower quadrant was shown by CT to be a Riedel's hepatic lobe. In several further patients coincidental lesions were documented in the CT report (gallstones 2, large renal cysts 2, fibroids 1, hepatic granulomata 1). DISCUSSION
(6) Fig. 3 - A n 85-year-old m a n who presented with melaena. (a) CT showing an irregular eccentric soft tissue mass distorting the caecum which has a very thick wall. The normal fat plane between the mass and the right anterior abdominal wall is absent; the abdominal wall muscles are ill-defined at this point. The features indicate anterior abdominal wall involvement which was confirmed at successful surgery. (b) Barium enema (BE) showing the carcinoma in the caecum. There were some difficulties in coating here; the bowel was also rather tortuous; BE classified as being of poor quality.
with a sigmoid colonic neoplasm causing a" block at BE, where proximal diverticular disease was only shown by CT. Otherwise there was agreement in the recognition of diverticular disease, although in four patients it appeared more extensive at BE than on CT. As regards other benign colonic disease, CT missed one case of Crohn's disease which was clearly demonstrated by BE. However, in retrospect, a segment of thickened bowel wall with
It appears that CT is easier to perform in these elderly patients than BE. Only in one patient was CT impossible (claustrophobia), whereas BE was unsuccessful at the first attempt in seven. In two of these seven patients it was considered that a second BE attempt would be unsuccessful. Thus a second BE was attempted in five (in one of which the necessity to obtain some BE images was certainly strengthened by the CT findings of a likely carcinoma in the ascending colon when the BE and CT findings were discussed). The fact that only 16 of the 33 BEs which were ultimately reviewed were considered of good quality confirms the known difficulty of performing BE in this age group [4]. Although the quality of the CT studies in this group was not formally assessed, we have reason to believe, from a previous study in this unit [10], that there is little deterioration in CT image quality with increased age. Nevertheless close attention to CT technique is important, particularly in the administration of oral contrast medium [8,11]. This is especially important when attempting to opacify the large bowel with prolonged oral contrast medium administration. Some might argue that we would have obtained better large bowel detail if we had administered air or contrast medium per rectum. However, that would have detracted from the advantages of a totally non-invasive test and led to some of the disagreeable features (incontinence, etc.) of the BE. In future it would probably be worth giving oral contrast medium the night before the CT examination. Interestingly, the patients did not object to BE all that much; only three of the 14 patients who found it unpleasant said they would never consent to it being done again. Furthermore CT only gained higher scores in 11 of the 21 comparisons using the visual scales. To our surprise two patients gave BE higher scores. However, there was a convincing overall preference amongst the patients for CT. Perhaps we should also have asked the radiologists, radiographer, nurses and other helpers for their opinions. CT fared well in the detection of colonic carcinoma, identifying all seven examples in this series. However, we appreciate that small lesions will be missed by CT and we
BARIUM ENEMA OR CT IN THE ELDERLY.9
fully support Balthazar's statements that 'barium examination remains superior to CT for evaluating intraluminal and mucosal disease' [8]. However, in the same article, the advantages of CT are described; 'CT is far more accurate for evaluating intramural and extraintestinal components'. As regards false positive findings, CT raised the possibility of a neoplasm in four patients in whom none was seen at BE. One (Fig. 2) was, in retrospect, an obvious example of the ileo-caecal valve pitfall. The difficulties of excluding tumour in and amongst diverticular disease, which caused problems in another patient, are well known [12]. Thus, if the CT report is anything other than definite for colonic carcinoma, BE or colonoscopy should be performed to further elucidate equivocally positive CT findings. A negative CT report (with regard to the colon) could be assumed to be correct unless there are very serious signs and symptoms. No significant lesion is found in quite a high proportion of these patients [1]. In this series it would have been safe to withhold BE in all those patients with negative CT except for the one with Crohn's disease. However, in retrospect, this diagnosis should have been considered on the CT images. The fact that CT can accurately determine the presence and extent of diverticular disease is well known [6]; the few incongruencies in our series probably reflects the wide increments that we used (2 cm) and the fact that we did not perform thin contiguous cuts in areas equivocal for diverticular disease. The extracolonic lesions identified by CT may provide the strongest case for recommending that CT should be the initial test in such patients. Involvement o f the abdominal wall (one patient, Fig. 3), involvement of the ureter and ascites (one patient), and the likely presence of hepatic metastases (two patients) are all findings which might influence the surgical approach. Indeed surgery was withheld in one of these patients and, in another, only a palliative defunctioning colostomy was performed; both died within 3 months. The significant gynaecological lesions (two in this series) were also well displayed by CT. In the early years of CT when the equipment was relatively Unsophisticated, its use in the gastrointestinal tract was mainly directed, quite properly, to the assessment of extraluminal disease. Accordingly it was recommended as a means of staging known tumours [13,14]. However, with increasing radiological experience, along with improved technical capabilities, its use has rapidly expanded and CT is now advocated as a diagnostic technique for many conditions in the gastrointestinal tract [8]. However, there is a wide range in the technical performance of the various CT machines presently operating in the UK. Furthermore many radiologists are still learning the more subtle CT features of large bowel lesions. We, with quite extensive CT experience, initially over-reported prominent ileo-caecal valves; we also failed to detect a segment of Crohn's disease which was quite obvious on retrospect. However, this study has made us pay even more attention to the large bowel and we would hope to do better in the future. Next come the questions of availability and cost. In many hospitals and communities CT remains a scarce facility. Accordingly the patients who might be referred for CT instead of BE may prevent other patients from being investigated by CT. This is largely a matter of
51
departmental organization. It may, logistically, be easier to schedule an additional abdominal CT than to look after a difficult patient in the BE screening suite. As regards cost, even a difficult BE (with additional nursing care) is likely to be just cheaper than CT for the radiological department. But if, by performing CT, a 2-3 day admission is avoided, the overall costs will be much lower if CT is used rather than BE. We propose, in the light of this study, to change our practice and use CT instead of BE as the initial test in these elderly patients suspected of harbouring large bowel disease; this policy will be kept under review. There will be a low threshold for recommending BE where CT is in any way equivocal or where the patient has continued unexplained symptoms. We may modify our CT technique to help recognize more subtle colonic lesions by administering oral contrast medium the night before CT and by reducing the increments between the slices. The original protocol was designed simply with the aim of detecting large carcinomas during a quick and inexpensive CT examination. We recommend that anyone considering a similar change would be well advised to also go through a period when both techniques are performed. We learnt a lot during this study. We hope that this policy of using CT rather than BE as the initial test for elderly patients suspected of having large bowel disease, which will be kept under review, will lead to less frustration, embarrassment and distress to all.
REFERENCES
1 Hill JC. Selection of elderly patients for barium enema examination with respect to significant bowel pathology especially carcinoma of the colon and rectum: the results of a retrospective study. Age and Ageing 1988;17:134-136. 2 Coni N, Davison W, Webster S. Lecture notes on geriatrics. Oxford: Blackwell Scientific Publications, 1988:203. 3 Stewart ET, Dodds WJ. Predictability of rectal incontinence on barium enema examination. American Journal of Roentgenology 1979;132:197-200. 4 Tinetti M, Stone L, Cooney L, Kepp MC. Inadequate barium enema in hospitalized elderly patients. Incidence and risk factors. Archives of Internal Medicine 1989;149:2014-2016. 5 Balthazar EJ, Megibow A J, Hulnick D, Naidich DP. Carcinoma of the colon: detection and preoperative staging by CT. American Journal of Roentgenology 1988;150:301-306. 6 Theoni RJ. CT evaluation of carcinomas of the colon and rectum. Radiologic Clinics o f North America 1989;27:731-741. 7 NeffCC, van Sonnenberg E. CT of diverticulitis. Radiologic Clinics of North America 1989;27:743-752. 8 Balthazar EJ. CT of the gastrointestinal tract: principles and interpretation. American Journal of Roentgenology 1991; 156:23-32. 9 Edwards RTM, Bransom C J, Crosby DL, Pathy MS. Colorectal carcinoma in the elderly: a geriatric and surgiCal practice compared. Age and Ageing 1983;12:256-262. 10 Baldwin J, Sharpe P, Cole S, Dixon AK. Image quality of abdominal computed tomography in the elderly. Age and Ageing 1987; 16:261264. 11 Dixon AK. Body CT." a handbook. Edinburgh: Churchill Livingstone, 1983:20-22. 12 Baker SR, Altermann DD. False-negative barium enema in patients with sigmoid cancer and coexistent diverticula. Gastrointestinal Radiology 1985; I0:171-173. 13 Dixon AK, Kelsey Fry I, Morson BC, Nicholls RJ, York Mason A. Pre-operative computed tomography of carcinoma of the rectum. British Journal of Radiology 1981;54:655-659. 14 Theoni RF, Moss AA, Schnyder P, Margulis AR. Staging of primary rectal and rectosigmoid tumours by computed tomography. Radiology 1981;141:135-138.
Barium Enema or Computed Tomography for the Frail Elderly Patient? J. J. DAY*, A. H. F R E E M A N , N. K. CONI* and A. K. D I X O N
Departments of Radiology and *Geriatric Medicine, Addenbrooke's Hospital and the University of Cambridge, Cambridge In order to determine whether abdominopelvic computed tomography (CT) offers an alternative to barium enema (BE) for the investigation of the large bowel in frail elderly patients, we have assessed and compared the results of both tests performed in each patient. Thirty-seven patients aged 71 to 88 (mean 80) with a history or clinical findings suggestive of large bowel disease were studied. The seven colonic neoplasms in this group were demonstrated by both techniques (apart from one patient who did not undergo BE as CT had shown an obstructing lesion). CT showed useful additional data in four patients (e.g. abdominal wall involvement) and demonstrated numerous extracolonic lesions (e.g. ovarian carcinoma). However, CT did raise the possibility of a large bowel neoplasm in four patients where none was shown by BE and missed one case of Crohn's disease. CT was the preferred test amongst the 25 patients where the acceptability of the two techniques could be compared. Only in 16 patients were the BE studies adjudged to be of good quality. CT should be the initial investigation of the large bowel in frail elderly patients requiring inpatient bowel preparation; the more unpleasant BE could be reserved for those cases where ~ T is equivocal or severe symptoms are unexplained. Day, J.J., Freeman, A.H., Coni, N.K. & Dixon, A.K. (1993). Clinical Radiology 48, 48-51. Barium Enema or Computed Tomography for the Frail Elderly Patient?
Accepted for Publication 21 January 1993
Barium enema (BE) is recognized as the standard radiological investigation for suspected large bowel disease. However, it is often a difficult procedure in elderly patients and each referral should be carefully considered [1]. Bowel preparation is frequently unsatisfactory and distressing. So much so that it is our usual practice (in common with many other geriatric units) to admit most of these patients for the preparation and the enema itself [2]. This adds to the expense of the investigation. Incontinence [3] and poor mobility may hinder the procedure. Furthermore, about one-third of BEs in this age group may be inconclusive due to the presence of faeces or incomplete demonstration of the whole of the large bowel [4]. Computed tomography (CT) has been shown to be capable of identifying most carcinomas of the large bowel [5,6], especially when these are large. CT can also demonstrate a wide range of mucosal abnormalities, such as diverticular and inflammatory bowel disease [7,8]. CT also provides additional information about other intra-abdominal structures which might be of value in these elderly patients, w h o often present with rather nonspecific symptoms and signs [9]. In order to see whether CT (a less invasive test) could replace BE in patients with features suggestive of large bowel disease, we performed CT on most patients who were referred for BE from the geriatric unit over a 5 month period. P A T I E N T S AND M E T H O D S Thirty-seven patients who were referred for BE from the Care of the Elderly Department at Addenbrooke's Correspondence to: Dr A. K. Dixon, University Department of Radiology, Addenbrooke's Hospital, Cambridge CB2 2QQ.
Hospital over a 5 month period were entered into the study. After obtaining verbal consent for an additional test to be done, CT was booked to coincide with the day of admission for the start of the bowel preparation (48 h before BE). On arrival the patients were given regular drinks of dilute oral contrast medium along with the laxative. Later that day, immediately before CT, they were given yet more oral contrast medium, the aim being to opacify the whole gastrointestinal tract. CT was performed using 1 cm thick slices at 2 cm increments through the whole abdomen and pelvis. In our unit, the decision as to whether to give intravenous contrast medium or to obtain further views (intervening slices, thin slices, decubitus images, delayed images, etc.) is up to the radiologist performing the CT examination; such additional procedures were performed as deemed appropriate. It is worth emphasizing that no patient received rectal contrast medium before CT. The CT study was reported, with special attention to the colon, by the registrar, senior registrar or consultant in charge of the session; registrars have to get their work checked by a more senior colleague. However, the CT report was withheld until the BE had been done. BE was carried out 2 days later, after 48 h of bowel preparation. A conventional double contrast technique was attempted wherever possible. The various manoeuvres available to facilitate complete opacification of the large bowel were used as appropriate. Again the procedure was performed by a registrar, senior registrar or a consultant; a registrar's report had to be checked by a more senior colleague. The report was made without access to the CT findings. The BE images were retrospectively reviewed and graded (AF) for quality into nondiagnostic, poor (allowing only a qualified report) or
BARIUM ENEMA OR CT IN THE ELDERLY.9
Fig. 1 - A 73-year-old man with anaemia, altered bowel habit and weight loss. CT shows clear evidence of small bowel dilatation culminating in a caecal mass. Note the thickened bowel wall around an irregular central area of contrast medium at the expected site of the caecum. This lesion was successfully removed at surgery. Barium enema was not performed.
good, according to the adequacy of the preparation, the degree of colonic visualization and the extent of barium retention. Following BE the results of both tests were made available and the management of the patients proceeded along usual lines. Only at this stage were the radiologists invited to comment on the two tests together (in cases of discrepancy). The patients went home as soon as possible after BE, with outpatient follow-up if necessary. Some patients remained in hospital for further treatment. The patients were asked by the referring clinician (within 3 weeks of the procedures) about their perception (pleasant/unpleasant) of the two techniques; these findings were analysed by the Chi-squared test. They were also asked to score each test on a 10 cm visual scale (where zero represented the worst experience ever and 10 the best); these scores were compared by the Wilcoxon's test for pair differences. At the end o f the study we analysed the radiological reports in each patient, along with the subsequent clinical course. Follow-up was arranged at 3 months, either through outpatients or the general practitioner so that any change in clinical condition could be assessed. In particular we classified the colonic findings into likely neoplasm, possible neoplasm, diverticular disease, inflammatory bowel disease and normal. RESULTS The age of the 37 patients ranged from 71 to 88 (mean 80) years. CT was possible in 36 o f the 37 patients; claustrophobia was the reason for the single failure. BE was technically impossible and completely abandoned in three patients (one at the second attempt). In four other patients the first BE attempt proved unsuccessful; images were only obtained at a second attempt. BE was not performed in one further patient as CT (Fig. 1) had shown a large caecal lesion which was causing small bowel obstruction. Amongst the 33 BEs which were available for review, two were deemed to be of non-diagnostic quality, 15 were judged to be of poor quality (although a qualified
49
Fig. 2 - An 86-year-old lady with constipation, abdominal pain and weight loss. CT shows a 2.4 cm mass (arrows) of apparent soft tissue in the caecum which also contains air and faeces. This was reported as a possible tumour. In retrospect it has the typical appearances of the normal ileo-caecal valve. Better opacification with oral contrast medium might have prevented this error of interpretation. The subsequent barium enema was normal with a normal ileo-caecal valve at this site.
conclusion could be made) and 16 were considered of good quality. Comparison of the acceptability of the two techniques was not possible in every patient because of dementia, illness, etc. Twenty-five patients were able to provide a simple comparison: 14 considered BE unpleasant (three said they would never consent to it being done again); CT was considered unpleasant by three (Z:= 10.75, P < 0.005). Twenty-one patients could give visual scores; these were higher for CT (mean 5.19 cm) than BE (mean 3.90 cm); 11 patients scored CT higher; eight patients gave both techniques the same score; two patients gave BE higher scores. The difference in these scores reached statistical significance (P < 0.05) on testing for pair differences. After 3 months follow-up, seven colonic (caecum 2, ascending colon 2, descending colon 1, rectosigmoid 2) neoplasms had been discovered amongst the 37 patients. All seven had been reported as a likely malignancy in the CT report; in one of these (Fig. 1), the presence of severe small bowel obstruction led to surgery without BE being performed. BE demonstrated the other six neoplasms, although in one patient the first BE had to be abandoned and the lesion in the ascending colon was only demonstrated on a second BE (in the knowledge of the positive CT findings). CT raised the possibility of a neoplasm in four further patients where no lesion was identified at BE. In one of these the radiologist over-reported the normal ileo-caecal valve (Fig. 2); in another there was concern about a possible lipoma of the ileo-caecal valve. In a third, the tissue planes between the rectum and vagina were indistinct; it transpired that the patient had previously received a caesium implant for carcinoma of the cervix. In the fourth patient a tumour could not be excluded in and amongst a segment of sigmoid colon affected by diverticular disease. There were no such false positive BE reports. Diverticular disease was mentioned in 19 of the 35 BE reports. On retrospective review of the BE and CT films there was quite close correlation in the recognition of diverticular disease between the two techniques. There were three patients where isolated areas of diverticular disease were only shown by BE. There was one patient,
50
CLINICAL RADIOLOGY
(a)
adjacent abnormal fat was visible on the CT images. In a patient with benign haemangiomatosis o f the rectum shown by CT the first BE was unsuccessful. The abnormality was only recognized in retrospect on the second BE. CT identified numerous extra colonic lesions. In two of the seven patients with colonic neoplasms CT suggested the presence o f hepatic metastases; in another, with a sigmoid carcinoma, CT showed ascites and an obstructed left ureter; in a fourth, CT showed abdominal wall involvement (Fig. 3). Amongst the other 30 patients CT revealed a large cystic ovarian mass in one patient in whom BE was impossible. In another, a complex pelvic paracolic mass, with air, fluid and calcium within, was demonstrated by CT; an infected gynaecological lesion was suggested; BE suggested a diverticular abscess or Crohn's disease; at surgery an infected ovarian dermoid was removed. In a further patient a possible mass in the right lower quadrant was shown by CT to be a Riedel's hepatic lobe. In several further patients coincidental lesions were documented in the CT report (gallstones 2, large renal cysts 2, fibroids 1, hepatic granulomata 1). DISCUSSION
(6) Fig. 3 - A n 85-year-old m a n who presented with melaena. (a) CT showing an irregular eccentric soft tissue mass distorting the caecum which has a very thick wall. The normal fat plane between the mass and the right anterior abdominal wall is absent; the abdominal wall muscles are ill-defined at this point. The features indicate anterior abdominal wall involvement which was confirmed at successful surgery. (b) Barium enema (BE) showing the carcinoma in the caecum. There were some difficulties in coating here; the bowel was also rather tortuous; BE classified as being of poor quality.
with a sigmoid colonic neoplasm causing a" block at BE, where proximal diverticular disease was only shown by CT. Otherwise there was agreement in the recognition of diverticular disease, although in four patients it appeared more extensive at BE than on CT. As regards other benign colonic disease, CT missed one case of Crohn's disease which was clearly demonstrated by BE. However, in retrospect, a segment of thickened bowel wall with
It appears that CT is easier to perform in these elderly patients than BE. Only in one patient was CT impossible (claustrophobia), whereas BE was unsuccessful at the first attempt in seven. In two of these seven patients it was considered that a second BE attempt would be unsuccessful. Thus a second BE was attempted in five (in one of which the necessity to obtain some BE images was certainly strengthened by the CT findings of a likely carcinoma in the ascending colon when the BE and CT findings were discussed). The fact that only 16 of the 33 BEs which were ultimately reviewed were considered of good quality confirms the known difficulty of performing BE in this age group [4]. Although the quality of the CT studies in this group was not formally assessed, we have reason to believe, from a previous study in this unit [10], that there is little deterioration in CT image quality with increased age. Nevertheless close attention to CT technique is important, particularly in the administration of oral contrast medium [8,11]. This is especially important when attempting to opacify the large bowel with prolonged oral contrast medium administration. Some might argue that we would have obtained better large bowel detail if we had administered air or contrast medium per rectum. However, that would have detracted from the advantages of a totally non-invasive test and led to some of the disagreeable features (incontinence, etc.) of the BE. In future it would probably be worth giving oral contrast medium the night before the CT examination. Interestingly, the patients did not object to BE all that much; only three of the 14 patients who found it unpleasant said they would never consent to it being done again. Furthermore CT only gained higher scores in 11 of the 21 comparisons using the visual scales. To our surprise two patients gave BE higher scores. However, there was a convincing overall preference amongst the patients for CT. Perhaps we should also have asked the radiologists, radiographer, nurses and other helpers for their opinions. CT fared well in the detection of colonic carcinoma, identifying all seven examples in this series. However, we appreciate that small lesions will be missed by CT and we
BARIUM ENEMA OR CT IN THE ELDERLY.9
fully support Balthazar's statements that 'barium examination remains superior to CT for evaluating intraluminal and mucosal disease' [8]. However, in the same article, the advantages of CT are described; 'CT is far more accurate for evaluating intramural and extraintestinal components'. As regards false positive findings, CT raised the possibility of a neoplasm in four patients in whom none was seen at BE. One (Fig. 2) was, in retrospect, an obvious example of the ileo-caecal valve pitfall. The difficulties of excluding tumour in and amongst diverticular disease, which caused problems in another patient, are well known [12]. Thus, if the CT report is anything other than definite for colonic carcinoma, BE or colonoscopy should be performed to further elucidate equivocally positive CT findings. A negative CT report (with regard to the colon) could be assumed to be correct unless there are very serious signs and symptoms. No significant lesion is found in quite a high proportion of these patients [1]. In this series it would have been safe to withhold BE in all those patients with negative CT except for the one with Crohn's disease. However, in retrospect, this diagnosis should have been considered on the CT images. The fact that CT can accurately determine the presence and extent of diverticular disease is well known [6]; the few incongruencies in our series probably reflects the wide increments that we used (2 cm) and the fact that we did not perform thin contiguous cuts in areas equivocal for diverticular disease. The extracolonic lesions identified by CT may provide the strongest case for recommending that CT should be the initial test in such patients. Involvement o f the abdominal wall (one patient, Fig. 3), involvement of the ureter and ascites (one patient), and the likely presence of hepatic metastases (two patients) are all findings which might influence the surgical approach. Indeed surgery was withheld in one of these patients and, in another, only a palliative defunctioning colostomy was performed; both died within 3 months. The significant gynaecological lesions (two in this series) were also well displayed by CT. In the early years of CT when the equipment was relatively Unsophisticated, its use in the gastrointestinal tract was mainly directed, quite properly, to the assessment of extraluminal disease. Accordingly it was recommended as a means of staging known tumours [13,14]. However, with increasing radiological experience, along with improved technical capabilities, its use has rapidly expanded and CT is now advocated as a diagnostic technique for many conditions in the gastrointestinal tract [8]. However, there is a wide range in the technical performance of the various CT machines presently operating in the UK. Furthermore many radiologists are still learning the more subtle CT features of large bowel lesions. We, with quite extensive CT experience, initially over-reported prominent ileo-caecal valves; we also failed to detect a segment of Crohn's disease which was quite obvious on retrospect. However, this study has made us pay even more attention to the large bowel and we would hope to do better in the future. Next come the questions of availability and cost. In many hospitals and communities CT remains a scarce facility. Accordingly the patients who might be referred for CT instead of BE may prevent other patients from being investigated by CT. This is largely a matter of
51
departmental organization. It may, logistically, be easier to schedule an additional abdominal CT than to look after a difficult patient in the BE screening suite. As regards cost, even a difficult BE (with additional nursing care) is likely to be just cheaper than CT for the radiological department. But if, by performing CT, a 2-3 day admission is avoided, the overall costs will be much lower if CT is used rather than BE. We propose, in the light of this study, to change our practice and use CT instead of BE as the initial test in these elderly patients suspected of harbouring large bowel disease; this policy will be kept under review. There will be a low threshold for recommending BE where CT is in any way equivocal or where the patient has continued unexplained symptoms. We may modify our CT technique to help recognize more subtle colonic lesions by administering oral contrast medium the night before CT and by reducing the increments between the slices. The original protocol was designed simply with the aim of detecting large carcinomas during a quick and inexpensive CT examination. We recommend that anyone considering a similar change would be well advised to also go through a period when both techniques are performed. We learnt a lot during this study. We hope that this policy of using CT rather than BE as the initial test for elderly patients suspected of having large bowel disease, which will be kept under review, will lead to less frustration, embarrassment and distress to all.
REFERENCES
1 Hill JC. Selection of elderly patients for barium enema examination with respect to significant bowel pathology especially carcinoma of the colon and rectum: the results of a retrospective study. Age and Ageing 1988;17:134-136. 2 Coni N, Davison W, Webster S. Lecture notes on geriatrics. Oxford: Blackwell Scientific Publications, 1988:203. 3 Stewart ET, Dodds WJ. Predictability of rectal incontinence on barium enema examination. American Journal of Roentgenology 1979;132:197-200. 4 Tinetti M, Stone L, Cooney L, Kepp MC. Inadequate barium enema in hospitalized elderly patients. Incidence and risk factors. Archives of Internal Medicine 1989;149:2014-2016. 5 Balthazar EJ, Megibow A J, Hulnick D, Naidich DP. Carcinoma of the colon: detection and preoperative staging by CT. American Journal of Roentgenology 1988;150:301-306. 6 Theoni RJ. CT evaluation of carcinomas of the colon and rectum. Radiologic Clinics o f North America 1989;27:731-741. 7 NeffCC, van Sonnenberg E. CT of diverticulitis. Radiologic Clinics of North America 1989;27:743-752. 8 Balthazar EJ. CT of the gastrointestinal tract: principles and interpretation. American Journal of Roentgenology 1991; 156:23-32. 9 Edwards RTM, Bransom C J, Crosby DL, Pathy MS. Colorectal carcinoma in the elderly: a geriatric and surgiCal practice compared. Age and Ageing 1983;12:256-262. 10 Baldwin J, Sharpe P, Cole S, Dixon AK. Image quality of abdominal computed tomography in the elderly. Age and Ageing 1987; 16:261264. 11 Dixon AK. Body CT." a handbook. Edinburgh: Churchill Livingstone, 1983:20-22. 12 Baker SR, Altermann DD. False-negative barium enema in patients with sigmoid cancer and coexistent diverticula. Gastrointestinal Radiology 1985; I0:171-173. 13 Dixon AK, Kelsey Fry I, Morson BC, Nicholls RJ, York Mason A. Pre-operative computed tomography of carcinoma of the rectum. British Journal of Radiology 1981;54:655-659. 14 Theoni RF, Moss AA, Schnyder P, Margulis AR. Staging of primary rectal and rectosigmoid tumours by computed tomography. Radiology 1981;141:135-138.