Protocol Online logo
Top : New Forum Archives (2009-): : Immunology

Calibration curve? - (Dec/11/2011 )


I have some problem dealing with 3 weeks till now with my ELISA kit for TSH and fT4. When i run new calibration curve the samples seem just fine, but when i run only samples and controls, they are not good, controls are out from the range and i cant be sure in my results. I have changed the kit, the temperature is just fine 20-25 degrees, i have changed the pippette, the water, the washer is ok it washes properly, so i dont know where i should be looking for the problem? Every time when i run a calibration curve the results are fine but should i run calibration each time for each sample or one calibration for one kit? The blank is fine. So i really need your answers from your experience here.

Thank you very much!


With manual methods you should run the entire dose response curve and controls each time you analyze samples. You may be able to run a few calibration points (L Mid and High) to do the analysis providing the curve shape is consistent.

Automated methods on the larger hospital analyzers 'store' the dose response curve data after the first run (which is stable for at least 30 days) and only samples and controls need to be run. This changes when new lots are used on the instrument and the entire curve has to be run.


Completely agree, you have to run your calibration curve every time.

In addition to that I have a question. When you say your samples are out of range, is that to the low or the high end? If it is the high end, you can try diluting your samples.

-almost a doctor-

I think we are troubleshooting a kit that a company made and sold. If this is so, then all problems should be directed to their technical support people.

With respect to high value samples > highest calibrator; they can be diluted but you must make certain that there is no matrix effect.


First of all tnx for your posts

The thing is that i have collegues that work on the same reader with the same kits and they do not have the same problem like me and they do not run calibration every time
@ almost a doctor: every time is different, sometimes they are lower and sometimes the oposite. For example TSH control is out of range to the lower end, and fT4 is out of range but to the higher end.
This confuses me because when i run new calibration, the samples are fine, but in a couple of days on the same calibration curve, which seems correct, i run the tests and i get some odd values. After that i have to run new calibration and after the new calibration the results are ok.
@ PAO_ahac: At first also thought that the problem was in the kit, so i have changed it, this is my third kit and the problem is still here. I don't think that the problem is in the kit, maybe in the reader or i'm doing something wrong, but i'm not runing ELISA for the first time. (this is my first time runing TSH and fT4 on ELISA). Also, i have noticed something that i think it should not be hapening. When i put stop solution, the absorbance should be stable at least 15 minutes, after stop solution when i read for the first time i have one value for the absorbance, but when i read 5 minutes later i have higher values for the absorbance, the reaction is not stable? But i have changed three kits and i have the same results.
Anyway i'm desperate so much that i really need to find a solution for my problem, otherwise, i have bought the reader for nothing.

p.s. tnx a lot nad have i nice day!!!


Ok. These are kits that you purchased. You should be able to discuss in person problems with the company. But, we will try to help.

1. You should run a curve each and every time you run the assay and not run too many samples as you may have 'end of run effect'. So do not do more that 1/3 plate.
2. Why does it make difference NOT to run calibrators each time...are you trying to save money by not using more wells? And, most importantly what is the name and manufacturer of the 'kit' and is their protocol/directions on-line. Let us know what the URL we can see what the instructions say!
3. On your first day everything is "fine"...well run the test that way all the time.
4. the color after adding stop will change slowy so just run your plate through once or twice and use ONE of the printouts. After you add the stop give a gently mix and read immediately that should be fine. then analyze your data at that time interval.
5 Run your controls with each and every run.
6. You said samples are do you know what the values of these samples are? Are they pre-analyzed by a reference method? Samples being 'fine' means that your results agree with the reference method.