Iron deficiency anaemia is a common problem both in general and hospital practice. The clinician is often presented with a patient where there are no clinical clues as to the source of the bleeding. The decision to know just how far to investigate these patients can be difficult especially in the elderly and infirm group of patients.

Occasionally a careful dietary history may be valuable but is usually only relevant in a small percentage of patients. Blood loss may be occurring outside the gastrointestinal tract and it is particularly important not to forget the urinary tract and those patients who are taking non-steroidal anti-inflammatory drugs or oral anticoagulants. Furthermore premenopausal women may not admit to excessive menstrual blood loss.

In this month's journal there are two very relevant articles addressing the problem of iron deficiency anaemia. The first is by Dr Jolobe and is primarily concerned with the definition and validation of iron deficiency and the appropriate investigations required to elucidate the underlying cause, particularly in elderly patients. The second paper by Willoughby and Litner retrospectively audits the investigation of iron deficiency anaemia in a district general hospital. It also suggests guidelines for future practice.

The latter is important in the age of protocols and best practice medicine as few hospitals are likely to have a policy for investigation of iron deficiency anaemia. These authors retrospectively studied the case notes of 334 patients endoscopically examined for anaemia. Altogether 126 were shown to have iron deficiency and a cause was identified in 48 (38%) patients.

Of the 48, four patients had a non-gastrointestinal cause including menorrhagia in two, dietary iron deficiency in one, and endometrial carcinoma in another. In the 44 who had a defined gastrointestinal cause, 15 patients had a colonic neoplasm, four had gastric cancer, and six other patients had coeliac disease. In 78 patients no cause for the anaemia was found and 74 patients were followed up for a median length of 28 months. Over this period menorrhagia was thought to be the cause in five patients and carcinoma of the caecum in a further one. Interestingly anaemia had resolved in 33 out of 74 patients.

In those patients who present with anaemia and who have no symptoms referable to either the upper or lower gastrointestinal tract, the investigation of choice well might be decided by the age of the patient. In younger patients an upper gastrointestinal endoscopy with duodenal biopsies looking for coeliac disease may well be more likely to yield a positive result while in the elderly patient investigation of the colon may be more important, particularly for the exclusion of colonic cancer.

Ideally the investigation of choice in the lower bowel is colonoscopy,1 although in a large number of hospitals the constraints of both expertise and facilities will result in most patients having a fibreoptic sigmoidoscopy and barium enema, or indeed a barium enema alone.

The classical catch is where a lesion of the upper gastrointestinal tract, which is thought to be the cause of anaemia, may well be an incidental finding and the patient is eventually found to have an occult neoplasm of the colon. Although this was not found in many patients in the Willoughby and Litner series, the percentage is likely to be much higher in the elderly population and figures of between 9% and 29% have been reported.2 It is therefore important, especially in the elderly, not to accept the finding of an upper gastrointestinal lesion as the cause of the anaemia without considering a more sinister cause in the colon.

A further problem arises in patients in which no abnormality is found after investigation of upper and lower gastrointestinal tract. The decision has to be taken whether it is worthwhile examining the small bowel. Enteroscopy may be very valuable particularly in patients who have been on non-steroidal anti-inflammatory drugs, but unfortunately is not available to most gastroenterologists.3 The diagnostic yield of barium examination of the small bowel is extremely disappointing,4 although in the Willoughby and Litner series two jejunal and two colonic cancers were diagnosed by this method.

What is particularly encouraging is that many patients where no definable lesion is found either in the upper or lower gastrointestinal tract seem to have a good prognosis, and in a high percentage the anaemia seems to resolve. This was certainly the case in the Willoughby and Litner series.

It may, however, be that patients seen in hospital outpatient departments who have their cases published in medical journals are very different from those seen in general practice who do not get referred to hospital. Many general practitioners sensibly may decide not to investigate their patients, particularly if they are elderly and infirm and just treat them with medication, that is, iron replacement.

A study with a long follow up looking into this group of patients in general practice who are not initially referred would be very interesting.

The message from both these papers emphasises the importance of making a definite diagnosis by taking a careful history not forgetting the possibility of extra gastrointestinal tract sources of blood loss. The possibility of dual pathology should always be considered and clinicians should not accept a minor abnormality in the upper gastrointestinal tract as a cause of the anaemia.