Anaemia: ferritin and mean cell volume helped diagnose iron-deficiency anaemia in elderly patients.

Clinical bottom line (level 1b)

  1. A third of elderly patients with anaemia had iron-deficiency anaemia.
  2. Physicians were fairly accurate at predicting patients's pre-test probabilities of iron-deficiency anaemia.
  3. A low ferritin in elderly patients with anaemia diagnosed iron-deficiency anaemia (LR+8.2) (for ferritin <45 mcg/l).
  4. A high ferritin in elderly patients with anaemia made iron-deficiency anaemia less likely (LR-0.13) (for ferritin >100 mcg/l).
  5. An MCV <74 made iron-deficiency anaemia more likely (LR+8.8) .
  6. An MCV >95 made iron-deficiency anaemia less likely (LR+0.11) .
  7. Other investigations were less helpful than ferritin or MCV.

Guyatt et al: American Journal of Medicine 1990; 88: 205-209
Patterson et al: Canadian Medical Association Journal 1991; 144 (4): 435-440

Expires November 2002

The study

Setting: two university-affiliated hospitals, Canada

235 patients (aged mean 80 years, 54% male) anaemia (Hb 12.0 g/dl or less in men; 11.0 g/dl or less in women, on at least two occasions)

Excluded if
  • <65 years old
  • institutionalised
  • recent blood transfusion or documented acute blood loss
  • too ill, impending death or severe dementia (based on physician's judgement)
  • bone marrow aspirate uninterpretable (9.3%)



    • Independent blinded reference standard, applied in all patients from a consecutive appropriate spectrum.
    Reference standard:
    • bone marrow aspiration by two histopathologists (K interobserver=0.84):
      • iron-deficiency anaemia if iron absent or decreased
      • anaemia of chronic disease: iron present in reticuloendothelial cells increased, and sideroblast numbers decreased
    Diagnostic test:
    • physician's prediction- based on history (anaemia, diet, bleeding including melena, ingestions of gastric-irritant drugs or other possible causes: weakness, fatigue, easy bruising, bone suggestive of myeloma or other malignant disease or chronic rheumatoid arthritis) and physical examination (abdominal mass, spontaneous bruising, lymphadenopathy, splenomegaly, melena or findings suggestive of chronic inflammatory disease)
    • ferritin
    • mean cell volume (MCV)
    • transferrin saturation
    • red cell protoporphyrin

    The evidence


    differential diagnosis number of patients prevalence
    (95% CI)
    final diagnosis: iron-deficiency anaemia 94 36%
    (30% to 42%)
    anaemia of chronic disease 113 44%
    (38% to 50%)
    megaloblastic anaemia 21 8.1%
    (4.8% to 11%)
    multiple myeloma 4 1.5%
    (0.043% to 3.1%)
    sideroblastic anaemia 3 1.2%
    (0.0% to 2.5%)
    dysmyeloplastic 3 1.2%
    (0.0% to 2.5%)
    other (eg. leukaemia, haemolytic anaemia, aplastic marrow, hypothyroidism, renal failure) 21 8.1%
    (4.8% to 11%)


    diagnostic test iron-deficiency-anaemia no iron-deficiency anaemia LR
    (95% CI)    		 post-test probability
    ferritin 18 µg/l or less 47 2 41
    (10 to 170)     		96%
    ferritin >18 to 45 or less 23 13 3.1
    (1.7 to 5.8)     		64%
    ferritin >45 to 100 or less 7 27 0.46
    (0.21 to 1.0)     		21%
    ferritin >100 8 108 0.13
    (0.067 to 0.25)     		7%
    total 85 150


    diagnostic test iron-deficiency anaemia no iron-deficiency anaemia LR
    (95% CI)    		 post-test probability
    transferrin saturation 5% or less 38 4 17
    (6.1 to 45)     		90%
    transferrin saturation >5% to 8% or less 14 17 1.4
    (0.74 to 2.8)     		45%
    transferrin saturation >8% to 21% or less 23 70 0.57
    (0.39 to 0.84)     		25%
    transferrin saturation >21% 9 55 0.28
    (0.15 to 0.55)     		14%
    total 84 146


    diagnostic test iron-deficiency anaemia no iron-deficiency anaemia LR
    (95% CI)    		 post-test probability
    MCV 74 µm³ or less 30 6 8.8
    (3.8 to 20)     		83%
    MCV >74 to 85 or less 32 42 1.3
    (0.92 to 2.0)     		43%
    MCV >85 to 91 or less 16 44 0.64
    (0.39 to 1.1)     		27%
    MCV >91 to 95 or less 5 26 0.34
    (0.14 to 0.85)     		16%
    MCV >95 2 32 0.11
    (0.027 to 0.45)     		6%
    total 85 150


    diagnostic test iron-deficiency anaemia no iron-deficiency anaemia LR
    (95% CI)    		 post-test probability
    red cell protoporphyrin >2 µg/l 40 24 3.0
    (1.9 to 4.6)     		63%
    red cell >1.25 to 2 or less 17 24 1.3
    (0.72 to 2.2)     		41%
    red cell >1.0 to 1.25 or less 9 21 0.77
    (0.37 to 1.6)     		30%
    red cell >0.75 to 1.0 or less 8 28 0.51
    (0.24 to 1.1)     		22%
    red cell 0.75 or less 10 53 0.34
    (0.18 to 0.63)     		16%
    total 85 150


    diagnostic test iron-deficiency anaemia no iron-deficiency anaemia LR
    (95% CI)    		 post-test probability
    physician's prediction >80% 10 1 18
    (2.3 to 135)     		91%
    physician's prediction >60% to 80% 9 2 7.9
    (1.8 to 36)     		82%
    physician's prediction >40% to 60% 18 7 4.5
    (2.0 to 10)     		72%
    physician's prediction >20% to 40% 4 10 0.70
    (0.23 to 2.2)     		29%
    physician's prediction 20% or less 30 105 0.50
    (0.38 to 0.67)     		22%
    total 71 125

    • Four geriatricians, four internists and two haematologists performed the predictions.

    Comments

    1. Results took a week to return in this study- most hospitals now provide a same-day service.

    Citation

    1. Guyatt GH, Patterson C, Ali M, et al: diagnosis of iron-deficiency anemia in the elderly. American Journal of Medicine 1990; 88: 205-209
    2. Patterson C, Guyatt GH, Singer J, et al: iron deficiency anemia in the elderly: the diagnostic process. Canadian Medical Association Journal 1991; 144 (4): 435-440
    Search Terms: from 'Practicing Evidence-based Medicine' by Strauss SE et al; publ. Radcliffe Press, 1998
    Contributor: Chris Ball and Clare Wotton, October 2000
    Reviewer:

    Clinical Question.
    Patient elderly patients with anaemia
    Intervention or Exposure physician's prediction, ferritin, MCV, transferrin, red cell protoporphyrin
    Comparison bone marrow aspiration
    Outcome diagnosis of iron-deficiency anaemia