Anaemia: cobalamin deficiency: a positive serum cobalamin assay did not diagnose.

Clinical bottom line (level 2b)

  1. Cobalamin deficiency was rare in patients suspected to have it.
  2. Patients with a gastrectomy were more likely to have cobalamin deficiency (LR+7.4) .
  3. Serum cobalamin assay can probably rule deficiency out, but did not help diagnose it (LR+2.1) .
  4. Urinary methylmalonic acid more than 5 µ g/mg creatinine made cobalamin deficiency much more likely (LR+47) .
Matchar et al: American Journal of Medical Science 1994; 308 (5): 276-283
Expires June 2003

The study

Setting: university hospital, USA

134 patients (aged mean 61 years, 99% male) in- or out-patients who had a positive serum cobalamin assay for anaemia (33%), macrocytosis (21%), dementia (15%), neuropathy (15%) (half with low assays; half with normal assays)

Excluded if
  • lived >1 hour's drive from hospital
  • unable to keep follow-up appointments
  • died before first follow-up evaluation



    • Independent blinded reference standard, applied in all patients from a non-consecutive appropriate spectrum.
    Reference standard:
    • cobalamin deficient if one of:
      • abnormal Schilling test (less than 7% excretion in 24 hours) and one of: Hb less than 13 g/dl that improved with cobalamin treatment; MCV more than 99 fl; mean neutrophil lobe count more than 3.6 per cell; macrocytosis on peripheral blood smear
      • normal Schilling test with MCV more than 99 fl, mean neutrophil lobe count more than 3.6 per cell or macrocytosis on peripheral blood smear that improved in response to parenteral cyanocobalamin
    Diagnostic test:
    • physical examination by blinded physician assistant
    • serum cobalamin (abnormal if less than 133 pmol/l)
    • spot urine measurement of methylmalonic acid (MMA) (abnormal if more than 5 microg/mg creatinine)

    The evidence

    pre-test probability of cobalamin deficiency: 12%, (95% CI: 6.4% to 17%)

    diagnostic test cobalamin deficiency no deficiency LR+
    (95% CI)    		 post-test probability LR-
    (95% CI)    		 post-test probability
    serum cobalamin <133 16 56 2.1
    (1.7 to 2.6)     		22% 0.0
    (0.0 to 0.33)     		0%
    history of gastrectomy 8 8 7.4
    (3.2 to 17)     		50% 0.54
    (0.33 to 0.88)     		7%
    total 16 118


    diagnostic test cobalamin deficiency no deficiency LR+
    (95% CI)    		 post-test probability LR-
    (95% CI)    		 post-test probability
    MMA >5 µg/mg creatinine 13 2 47
    (12 to 190)     		87% 0.19
    (0.07 to 0.53)     		3%
    total 16 115

    • No other clinical signs were found to be significant indicators of cobalamin deficiency.

    Comments

    1. 1599 patients were originally screened- positive test rate 5.2%.
    2. 133 pmol/l was found to be the cut-off point that maximised sensitivity whilst maintaining specificity.

    Citation

    1. Matchar DB, McCrory DC, Millington DS, et al: Performance of the serum cobalamin assay for diagnosis of cobalamin deficiency. American Journal of Medical Science 1994; 308 (5): 276-283
    Search Terms: cobalamin in Cochrane
    Contributor: Chris Ball and Clare Wotton, June 2000
    Reviewer:

    Clinical Question.
    Patient positive serum cobalamin assay
    Intervention or Exposure physical examination, serum cobalamin, measurement of methylmalonic acid
    Comparison abnormal Schilling test
    Outcome diagnosis of cobalamin deficiency