No matter what field you are in, if you run an instrument that measures anything—the speed of a car on a radar gun, the heat of an oven that bakes perfect pizzas or, in our case, the amount of cholesterol (or other analytes) in a vial of serum; you need to do a few things to “prove” you are measuring what you think you are measuring correctly.
Calibrators: A calibrator does what it says—it calibrates an instrument. As a researcher in Germany, I used to set up experiments that measured how much of an analyte I had in a sample of pig’s brain (among other things). The first thing I had to do was to set up a calibration curve for my measuring instrument. I used to have three points of known analyte concentrations—why three? Well, I wanted an upper limit and a lower limit and somewhere in the middle. That way I could plot all three on a graph with readings along the vertical axis and amount of protein along the horizontal axis and hopefully get a straight line going through the readings.
I would usually get something like this: 
- If you add a trendline, the data points will make a straight line (sort of).
- If you extend the line down to the “X” axis, it will not go through zero but more like 75 or so.
What can we learn from this? Well, not all readings, even with the best care and set up, will line up perfectly and even when they do, it is rare that they will do so all the way to zero. But still, if I keep the “unknowns” I am reading between say 200 and 450 on my reading scale, I will be close to getting the right answer.
So, does it surprise you that most modern clinical chemistry analyzers only use one calibrator to set their machines? The second reading is assumed to be zero! If we do this and use the high x, we get the graph below. Something to think about.
OK—but if you are a clinical chemist or technician, you will want a nice high calibrator of the analyte you are looking to measure. Then all you must do is keep your concentrations in a range to give you readings of the calibrator and zero (preferably closer to the calibrator than to zero).
So—a calibrator is a solution that will give you a high reading of an analyte you want to measure. You use it to set the upper reading you can use in your work and in the case of most modern clinical chemistry analyzers will assume that the zero reading is OK.
Controls: A control is a sample with a known value that you run with your patients or donors’ samples. You now have an internal check of how close your unknown samples are to being correct when you compare your control readings with what they should be. If the controls all pass you can call your run, “in control.” If you go to the official US government CLIA (Clinical Laboratory Improvement Amendment-88) website, it will give you a list of the analytes in most laboratories and what range of errors you are allowed to make when running patients or donors etc. for these different analytes. There are also rules on how many controls to run and how often etc.
Proficiency Testing: Still, with your own calibrators and your own controls, the government doesn’t completely trust you to give out the right answers (the power of the profit motive). So, in the US if you are running donor or patient serum or plasma or urine and report analyte values to them or their doctors for diagnostic purposes etc. you must also pass relevant proficiency surveys or proficiency tests. In these tests an outside organization sends you samples that are analyzed elsewhere and you test the samples as unknowns and submit your values to the agency that is testing you.
One of the barriers to standardizing results from different laboratories has been the variety of instruments and reagents that different laboratories use to generate patient results.
Early on, this forced proficiency testing and regulatory agencies to establish procedures to judge a laboratory’s results. One obvious method of comparison was to establish peer groups among laboratories being tested. The user of a particular instrument would be judged against other laboratories using the same instruments and reagents. However, laboratories with older or esoteric instruments would be graded with very small peer groups, sometimes only themselves, which made them essentially not graded. In addition, the differences between peer groups could vary widely depending on the analyte being tested. That is an individual tested on one instrument could vary by a significant amount when compared to that same individual being tested on another instrument.
Essentially, two things needed to happen before these surveys could be compared across all instruments and reagents.
Reference Procedures: First, there needed to be a clear method (a gold standard for setting a test’s value) for any analyte that would give a value that transcended any instrument or reagent. There are a few laboratories for some analytes that are certified as reference sites. A good example of this is the CRMLN (Cholesterol Reference Method Laboratory Network) laboratory system monitored by the Centers for Disease Control and Prevention.
Commutable Source: Second, a commutable source of material that could be measured by all instruments and get the same result needed to be developed. Commutable simply means that all instruments or methods used to measure this analyte would get the same result when using the commutable source. This was provided for some time by the simple procedure of sending out fresh patient samples to laboratories by overnight or local delivery services. A good example of this would be the CDC’s Cholesterol Certification Program where 6 fresh patient samples exhibiting a wide range of cholesterol values were sent to laboratories and the analyses were graded by the CDC for each laboratory against the reference values set at the CDC by the reference procedure.
There were many practical problems with this method, most important being that only a limited number of laboratories could be serviced at any one time with fresh material. The only other source of samples for proficiency testing was lyophilized serum (serum that has been freeze-dried then rehydrated by the laboratory before use) which presented many problems such as mixing, solubility and correctly adding the specified amount of water to the lyophilized sample.
To solve these problems, in the late 1990s the CDC and Solomon Park Research Laboratories and several other agencies produced two pools with high and low levels of cholesterol processed by the method described in the CLSI C37-A procedure. Materials produced by this method are now used for the CDC’s LSP (Lipid Standardization Program) which evaluates approximately 100 laboratories worldwide. Serum produced by the C37-A procedure have also been used in several studies and reported in numerous journals.
Serum produced by the C37-A procedure is available at Solomon Park Research Laboratories.
If you have any questions, contact us today and we will be happy to answer.