Diagnosis of Iron-Deficiency Anemia in the Elderly
The American Journal of Medicine Vol. 88 March 1990
Authors: Gordon H. Guyatt MD, Christopher Patterson MD, Mahmoud Ali MD, Joel Singer MD, et al
Background
*Iron deficiency anemia is the second most common cause of anemia in the elderly
*Appropriate replacement therapy may improve symptoms.
*Herald occult gastrointestinal pathology
*Inappropriate Iron therapy may cause unwanted side effects
*Bone Marrow Biopsy is the gold standard for diagnosis of Iron deficiency anemia
* Pioneer study to apply multiple cut-points for a lab test to generate likelihood ratios
Clinical Question: What is the value of serum ferritin, mean cell volume, transferrin saturation, and free erythrocyte protoporphyrin in the diagnosis of iron-deficiency in the elderly?
Study Type: Prospective study, blinded at two hospital centers in Ontario, Canada
Study Population:
Mean Age: 79.7 +/- 7.62 years; 46% men; Mean Hgb 9.81g/dL; The most common medical diagnoses in this study group aside from anemia are early dementia, CHF, chronic airflow limitation, rheumatoid arthritis, osteoarthritis, pneumonia.
Inclusion Criteria:
-- > 65 years of age
-- Men; Hgb < 12.0g/dL on two separate occasions, Women; Hgb < 11.0 g/dL
Exclusion Criteria:
--Patient judged to ill, demented, or terminal
--Refused consent for bone marrow aspiration
--No evidence of anemia on second Hgb determination
--Recent transfusion (Hospital with busy ortho service with many post-op hip fractures)
--Previous bone marrow aspiration had revealed the diagnosis
--Institutionalized
Methods:
A total of 1,334 patients over 65 year of age were identified
Based on exclusion criteria above, 259 patients remain
The 259 patients underwent bone marrow biopsy, of these results 235 were interpretable
The bone marrow biopsies were performed and interpreted by a hematologist who was unaware of the results of the lab tests.
The results of the biopsies were classified as iron absent, reduced, present or increased.
After interpreting the results, the hematologist reviewed all relevant clinical information and made a final decision regarding the cause of anemia.
Results of the first 65 marrow aspirations were also interpreted by a second hematologist (also unaware of the lab findings) and discrepancies resolved by consensus.
The weighted kappa-quantifying chance corrected score for the 65 marrow aspirates interpreted by two hematologists was 0.84.
Results:
Receiver Operating Characteristic (ROC) curves were calculated for each of the 4 diagnostic tests
p.207 Figures 1-4
Statistical Analysis revealed the area under the ROC curves was 0.91(ferritin), 0.79(transferrin sat), 0.78 (MCV) and 0.72 (RCP). The difference between ferritin and each of the other three tests was statistically significant (p < 0.001 in each case)
*An area of 1 represents a perfect test; an area of .5 represents a worthless test.
Likelihood ratios for the four tests are represented
p.208 Table III
Therefore, ferritin proved to be a powerful predictor in determining iron deficiency anemia from anemia of chronic disease in the elderly.
36% of the patients in this study had no demonstrable iron in the marrow and were classified as iron deficient. Table II
The likelihood ratios are as follows: Ferritin level > 100ug/L, 0.13; greater than 45 ug/L but less than or equal to100 ug/L, 0.46; greater than 18 ug/L but less than or equal to 45 ug/L, 3.12; and less than or equal to 18ug/L, 41.47. These results indicate the values up to 45 increase the likelihood of iron deficiency, whereas values over 45 decrease the likelihood of iron deficiency.
In ROC curves: the TP rate (sensitivity) is plotted along the FP rate (1-specificity) for different cut off points. Each point on the ROC plot represents a sensitivity/specificity pair corresponding to a particular decision threshold.
Likelihood Ratio: The probability of a given test result among people with a disease divided by the probability of that test result among people without the disease.
+LR: Likelihood of + test with disease = A/(A+C) = SN/1-SP
Likelihood of + test without dz B/ (B+D)
-LR: Likelihood of – test with disease = C/ (A+C) = 1-SN/SP
Likelihood of – test without dz
In order to calculate LR from the ROC curves it is simply the slope of the TP/FP rate for any given cutoff point.
Most laboratory tests are reported on a numerical scale -- not merely as normal or abnormal. Thanks to ROC curves and likelihood ratios we can better understand how significant these data points are in a given range. By grouping numerical lab values into cutoff points determined by sensitivity and specificity we can then determine probability of disease to then guide our therapeutic management.
Are the results valid?
1. Oftentimes with Iron Deficiency Anemia there is diagnostic uncertainty based on clinical history and exam. The test of ferritin with a strongly +LR at values < 18 can help confidently generate probability of the disease. In other words, this test even demonstrates usefulness with intermediate pre-test probability. Thus, it has the power to decrease your action threshold for treatment.
2. The investigators tested appropriate patients geared to answer the clinical question. Elderly patient > 65 years of age and also using the definition of anemia based on Hgb levels. Exclusion criteria also appeared reasonable and appropriate.
3. Are my patients similar to those in the study? Yes, tertiary care elderly patients with a variety of co-morbid illnesses. We can not extrapolate this data to our younger population in the scope of this paper.
4. The gold standard is accepted in Iron Deficiency anemia as the bone marrow biopsy. However this is a very invasive test which necessitates promoting a less invasive although accurate test.