Lesson of the week: prevalence of concomitant disease in patients with iron deficiency anaemiaBMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7075.206 (Published 18 January 1997) Cite this as: BMJ 1997;314:206
- Simon H Till, registrar in general medicine and rheumatologya,
- Michael J Grundman, consultant physician and gastroenterologista
- a Department of Gastroenterology Chesterfield and North Derbyshire Royal Hospital Chesterfield S44 5BL
- Correspondence to: Dr Grundman
- Accepted 19 September 1996
How to investigate patients presenting with iron deficiency anaemia is a matter of debate. Undoubtedly one of the commonest lethal, but potentially curable, causes is colonic carcinoma, the second most common cancer in both men and women in the West. Despite this, the colon may not be investigated if an acceptable cause for the anaemia has been found on upper gastrointestinal investigations.
Clinically important concomitant disease, such as benign upper gastrointestinal disease associated with colonic cancer, has been reported in up to 7% of anaemic patients,1 2 3 yet combined upper and lower gastrointestinal investigations are carried out in only a minority of cases with iron deficiency anaemia.4
Patients, methods, and results
We examined the records of 89 consecutive patients referred with iron deficiency anaemia and at least two positive faecal occult blood results between 1989 and 1992. Iron deficiency anaemia was defined by a haemoglobin concentration of <110 g/l in association with a mean corpuscular volume of <80 fl and either a serum ferritin concentration of <10 mg/l or an appropriate response to iron supplementation. Patients were either followed up or their case notes were reviewed or their general practitioner was contacted. The case notes of all patients were reviewed in July 1995, a minimum of three years after initial presentation. We identified “acceptable” causes for the anaemia in 67 patients (table 1)). Six patients had concomitant disease; in three potentially curative surgery was delayed because of the gastroscopy findings (box). Cases 1-7-and a more recent case (8)-show the importance of not ascribing the cause of iron deficiency anaemia to a lesion in the oesophagus, stomach, or duodenum unless the colon has been investigated.
Concomitant disease in eight cases of iron deficiency anaemia
Case 1-65 year old man; haemoglobin 83 g/l; no gastrointestinal symptoms. Deep Barrett's ulcer shown by gastroscopy. Iron and ranitidine resolved ulcer and anaemia. Latter recurred after stopping treatment. Barium enema six months later showed caecal mass. Right hemicolectomy for Dukes' C caecal carcinoma.
Case 2-67 year old man; increasing dysphagia; haemoglobin 68 g/l. Previous benign oesophageal stricture. Treated with iron and omeprazole for severe oesophagitis. Anaemia recurred 10 months later; barium enema indicated caecal mass. Palliative hemicolectomy for annular Dukes' C caecal carcinoma.
Case 3-72 year old man; haemoglobin 77 g/l; minimal fresh rectal bleeding. Erosive gastritis at gastroscopy and caecal mass on barium enema. Right hemicolectomy for Dukes' B caecal carcinoma six weeks later.
Case 4-79 year old woman; haemoglobin 107 g/l; no gastrointestinal symptoms. Erosive gastritis at gastroscopy and “apple core” lesion in ascending colon on barium enema. Right hemicolectomy for Dukes' B carcinoma one month later.
Case 5-69 year old man; haemoglobin 104 g/l; epigastric discomfort. Grade 1 oesophagitis and ascending colon filling defect. Right hemicolectomy for Dukes' B carcinoma.
Case 6-75 year old man; haemoglobin 105 g/l; emergency admission; history of haematemesis, melaena, iron deficiency anaemia. Oesophageal ulcer treated with omeprazole. Barium enema for recurrent anaemia showed ascending colonic mass; Dukes' C carcinoma confirmed at surgery.
Case 7-70 year old woman; abdominal mass confirmed as Dukes' B caecal carcinoma. Investigated seven years earlier for iron deficiency anaemia: normal results; iron treatment resolved anaemia. This recurred three years later; gastroscopy and colonoscopy to hepatic flexure only were normal. Repeat barium enema advised but declined by patient.
Case 8-68 year old man; reflux helped by ranitidine; haemoglobin 67 g/l. Chronic duodenal ulcer on gastroscopy and mass at hepatic flexure on barium enema examination. Right hemicolectomy for Dukes' B carcinoma two weeks later. (Note: this case is not included in the original series of 89 patients.)
The prevalence of colonic cancer among patients presenting with iron deficiency anaemia varies considerably, with outpatient studies suggesting rates of 4-11%.1 3 5 6 We diagnosed malignant disease of the colon in 14 (15%) patients. Of the 13 patients who were initially found to have colonic cancer, six (46%) had “acceptable” upper gastrointestinal causes for their anaemia. The incidence of colonic cancer increases with age, 94% occurring in patients over the age of 50.7 All our patients were aged over 60, and all cancers were proximal to the splenic flexure. The higher prevalence of colonic cancer in our audit might be attributed to our use of faecal occult blood loss as a selection criterion, in contrast to other studies,1 3 5 6 but the role of faecal occult blood testing in iron deficiency remains to be evaluated.
The risks of not investigating the colon have been further emphasised by a recent audit of the investigation and outcome of iron deficiency anaemia.4 Long term follow up identified two patients (at 21 and 18 months) with advanced carcinoma of the colon. Neither patient had had colonic investigations at the time of initial presentation.
Of the six patients with concomitant disease, three had potentially curative surgery delayed because of their gastroscopy findings (cases 1, 2, and 6). The delay in the cases 2 and 6 was particularly lengthy (10 and 12 months, respectively). We suspect that had colonic investigations been carried out at the time of initial presentation, the diagnosis would almost certainly have been made.
The potential for delay in diagnosing proximal colonic cancer was first raised in 1969.8 Fagan reported that 10 out of 96 cancers of the right side of the colon were initially misdiagnosed as an upper gastrointestinal lesion on the basis of a positive finding on barium meal examination. A more recent study concluded that a delay in referral for investigation was the main avoidable reason for delay in diagnosis rather than the finding of an upper gastrointestinal lesion.9
Symptoms did not correlate well with the final diagnosis. This has been shown before.1 2 3 10 Of the six patients with concomitant disease, three had upper gastrointestinal symptoms while none had proximal colonic symptoms. Only Rockey et al reported that colonic symptoms predicted colonic disease.5
Follow up of patients in 1995, at least three years after their initial investigations, has not shown serious gastrointestinal disease in patients discharged after negative results on gastroscopy, duodenal biopsy, and barium enema, confirming that this is a safe, limited approach to the investigation of patients presenting with iron deficiency anaemia.11 Duodenal biopsy is a simple process during gastroscopy, and many series have shown a prevalence of coeliac disease in iron deficiency anaemia of 3-5%.1 One patient (case 7) had developed a Dukes' B ascending colon carcinoma at follow up. A barium enema had originally been advised but declined by the patient.
Although colonoscopy is accepted as the best way to visualise the colon,12 less experienced workers may fail to reach the caecum in up to half the cases.13 14 15 Double contrast barium enema may occasionally miss colonic cancer, but these are usually in the sigmoid.16 We found that colonic cancers presenting with iron deficiency anaemia were proximal to the splenic flexure, suggesting that double contrast barium enema is the investigation of choice, with colonoscopy reserved for persistent or recurrent anaemia or when doubts have been raised by the results of barium enema examination.
We believe that the risk of missing colonic cancer in patients with iron deficiency anaemia is sufficient to justify colonic examination, if findings at gastroscopy are benign. A cut off age of 40 would seem appropriate in view of the evidence that almost all colonic cancers occur over the age of 50.7 Younger patients without a family history or predisposing condition should probably have colonic investigations for persistent anaemia.
We thank the department of clinical audit at Chesterfield Royal Hospital for their help. We are particularly grateful to Christina Fielding for her advice during the planning of this study and Jan Woodward for administrative support.
Some of these data were presented at the British Society of Gastroenterology meeting in 1992 and published in abstract form in Gut 1992;33: S31.