Introduction
Here I consider the importance of two sets of figures that appear in pathology reports:
 Reference Intervals
 Limits, or Cutoffs
I shall use the three analytes of particular interest to me, for the diagnosis of vitamin B12 deficiency, as examples of the use of incorrect limits:
 Vitamin B12
 Methylmalonic Acid
 Total Homocysteine
Quoting incorrect reference intervals or limits can significantly affect the interpretation of results. In my case, investigating vitamin B12 deficiency, the use of incorrect limits caused an actual deficiency to be reported as a normal condition.
Reference Intervals
A reference interval is the range of values of a test result for a defined population. The range is usually specified as the central 95% of the values, from the lower 2.5% to the upper 97.5% of the population. In a normal distribution, 50% of the population will have a result below the mean, and 50 % will have a result above the mean, as shown in the following diagram:
What the Reference Interval Does Show
The reference interval shows:
 How broad the range of results is for a specific lab, for a given test, for a given population
 How the result for a patient compares with the rest of the population
What the Reference Interval Does Not Show
 What result is likely to indicate disease
 What range of results is indeterminate for indication of disease
 What range is unlikely to indicate disease
So, the reference interval tells us how our result compares to the rest of the population, but does not tell us if the result is a healthy one.
For example, my weight and blood pressure are very much lower than the average Australian male of my age. With 60% of men overweight or obese, is quite possible that I am in the lower 2.5% of the population for both of these measures, but this, by itself,says nothing about my state of health.
What we need to know is what is a healthy range; this is why limits, or cutoffs, are important.
There are some excellent articles and slide shows on the subject of reference intervals, including Reference Intervals for New Methods, reference AB24, and Statistical Techniques for Reference Intervals, reference AB25
Limits
A limit, or cutoff is a value that indicates a change from a result that is not indicative of disease to one that is indicative of disease. The important points about limits are:
 Their level is determined by research into the relationship between the analyte and the disease
 They are independent of the reference interval
 They are absolute values and are independent of the lab or the method of analysis
 They are only meaningful if the analytical method is sufficiently accurate
For some analytes, for example methylmalonic acid and homocysteine, there is a single cutoff point for a given disease, although the actual level might not be universally agreed to:
For other analytes, for example vitamin B12, there is a range of indeterminate values:
Because limits are taken from research, and compare the patient's result to absolute values, accuracy is critically important. This is explained well in the ERNDIM article, reference AB01:
Not only is accuracy of the result important; so also is the use of the correct cutoff in the pathology report. Where inappropriate limits are quoted, the interpretation of the result can be very significantly affected.
The labs cited here are not alone in quoting incorrect cutoffs; they also appear in the RCPA Manual.
Top of page
