Anaemia: iron-deficiency: lack of symptoms did not exclude upper and lower GI disease in elderly patients - further investigations were required.

Clinical bottom line (level 4)

  1. The commonest causes of iron-deficiency anaemia in the elderly were upper GI lesions - mainly ulcerative disease. The commonest lower GI cause was colonic neoplasia. No cause could be found in a seventh of patients.
  2. Elderly patients with iron-deficiency anaemia and upper GI symptoms were more likely to have an upper GI lesion (LR+4.2) . Absence of symptoms made upper GI lesions less likely but further tests were required.
  3. Lower GI symptoms were not very helpful at diagnosing or excluding lower GI lesions.
Cook et al: British Medical Journal 1986; 292: 1380-1382
Expires October 2003

The study

Setting: acute hospital, Australia

100 patients (aged mean 70, 69% male) with iron deficiency anaemia (Hb < 110 g/l and MCV < 80 fl, and transferrin <5% saturated, or low ferratin, or reduced iron on bone marrow smear, and response to iron therapy)

Excluded if
  • aged > 85
  • general debility precluded investigation
  • known cause for blood loss (surgery, trauma, overt GI bleeding)
  • premenopausal women
  • established reason for anaemia: renal failure requiring haemodialysis or a haemoglobinopathy
  • recently fully assessed for iron-deficiency anaemia
  • confirmed malignancy at time of admission



    • Independent unblinded reference standard, applied in all patients from a consecutive appropriate spectrum.
    Reference standard:
    • Diagnoses were made using the following investigations
      • gastroscopy (100%)
      • colonoscopy or air barium enema and flexible sigmoidoscopy (90%)
      • follow-up for at least 2 months if normal investigations or no colonic assessment (mean follow-up 6 months)
    Diagnostic test:
    • upper GI symptoms
    • colonic symptoms

    The evidence

    pre-test probability of upper GI disease: 60%, (95% CI: 50% to 70%)
    pre-test probability of lower GI disease: 30%, (95% CI: 21% to 40%)

    differential diagnosis number of patients prevalence
    (95% CI)
    ulcerative upper GI lesions 36 36%
    (27% to 45%)
    upper GI malignancy 6 6.0%
    (1.3% to 11%)
    hiatus hernia (>10 cm) 7 7.0%
    (2.0% to 12%)
    oesophageal varices 3 3.0%
    (0.0% to 6.3%)
    angiodysplasia 5 5.0%
    (0.7% to 9.3%)
    gastric surgery 10 10%
    (4.1% to 16%)
    colorectal neoplasia 20 20%
    (12% to 28%)
    miscellaneous 2 2.0%
    (0.0% to 4.7%)
    no cause found 14 14%
    (7.2% to 21%)


    diagnostic test upper GI disease no upper GI disease LR+
    (95% CI)    		 post-test probability LR-
    (95% CI)    		 post-test probability
    upper GI symptoms 31 5 4.1
    (1.8 to 9.6)     		86% 0.49
    (0.35 to 0.69)     		43%
    total 54 36


    diagnostic test lower GI disease no lower GI disease LR+
    (95% CI)    		 post-test probability LR-
    (95% CI)    		 post-test probability
    lower GI symptoms 11 12 2.1
    (1.1 to 4.2)     		48% 0.73
    (0.52 to 1.0)     		24%
    total 27 63

    Comments

    1. Note that 7/44 patients (16%: 95% CI: 5.1% to 27%) with benign upper GI lesions had colonic carcinoma or adenomas.
    2. No cause could initially found in 21% of patients - three were found subsequently to have a cause (GI erosions and two bladder cancers).

    Citation

    1. Cook IJ, Pavli P, Riley JW, et al: Gastrointestinal investigation of iron-deficiency anaemia. British Medical Journal 1986; 292: 1380-1382
    Search Terms: anaem* or anem* in Cochrane
    Contributor: Chris Ball and Bob Phillips, October 1999
    Reviewer:

    Clinical Question.
    Patient anaemia
    Intervention or Exposure GI symptoms
    Outcome GI lesions