To make things simple, if the 5 mins you want to remove is at the end of the gradient, or the slope of the gradient does not change until the target elutes, then the LOQ/LOD and other things do not change. If the slope of the gradient is changed, saying the peaks are flattened or squeezed, then the LOQ will change.

If you use Isocratic run, it may be ok. But if you use gradient, the column equilibration may change. When the last peak come out, you need to ramp up the organic to clean more non-polar interference and equilbrate to the initial condition long enough to ensure that it reaches the point where you want to start, e.g. 90% aqueous. There is a delay between the pump and the column. For instance, you program to have 100%B at the pump,  it will take a while before the end of the column will see 100% B.  

You can easily check the LOQ by running both methods (in matrix) and compare the chromatograms to see if there is a significant difference.  You also can check retention time reproducibility at the same time. It may not affect % recovery.

Trevor and Narong make very valid comments, you should balance the improved throughput against the effort required to be sure the adjusted method is giving the same results. Are there mycotoxins analysed in the method that you aren't looking for? that may be another explanation for the longer run time other than those already suggested around column cleanup and re-equilibration prior to the next run.

You must at least do an equivalency study using the matrix or matrices in which the analytes are. It really is peace of mind for you. You could also look for column life, back pressure and background in your comparison.

One question you must also ask yourself is whether it is an official method of analysis and you are producing COA for clients, in which case the equivalency study will need to be documented and submitted before you can produce COA documents for clients.

All the explanation given by above experts are right and should be taken into consideration before analysis. I agree with all the comments. 

Just a few points which may help understand what the other have said.

You wrote that "we want to remove 5min and set the total run time to 20min". Narong already wrote that this could be valid for an isocratic run. And I agree with that.

So far you do not modify MP composition and you peaks RT is unvaried, then you should not see variation of LoD, precision etc etc.

A simple re-run of LoQs and LoDs replicates should be sufficient to prove it.

From what you wrote, I have the feeling this is not a gradient run. If it is, then Trevor and Narong comments are all you need.

Just to make it clear: if you modification of the LC program results in 

- different RT

- different peak shape

Then, you need to re-evaluate all the validation parameters.

To make things simple, if the 5 mins you want to remove is at the end of the gradient, or the slope of the gradient does not change until the target elutes, then the LOQ/LOD and other things do not change. If the slope of the gradient is changed, saying the peaks are flattened or squeezed, then the LOQ will change.

A small addition to what Yansheng wrote (which is exactly what I meant).

If you cut part of the high organic solvent washing sector or the re-equilibration sector, you will most likely not see changes in LoQ or any other parameters for a few runs.

You will notice changes in column stability and response over several injections.

If you have inter-day reproducibility among your parameters, then please take this into account. 

Hello. I think of your question all is said above. In gradient elution systems, not only the proportion of solvents is important but the rate at wich the change happens also.

More idea, if say your last peak is eluted at 12 min and you want to shorten time, during the equilibration step at the end of the run, you can increase the flow rate so that the total volume of the mobile phase of the initial condition is the same e.g. the origianal method is 0.5 mL/min for 5 min (2.5 mL), you can do 1 mL /min for 2.5 min or 1.25 mL/min for 2 min. This way, there is no change in chromatography. You can also increase the flow rate after the last peak is out to wash the column with the highest organic, same idea, it will shorten the time without effecting the rest of the parameters.

If you are removing run time from the end of the run (when the column is washing or equlibrating to the starting conditions) you will not observe changes in LOD/LOD, recovery or precision. You may test your results by analyzing samples using old  vs new method. But if you planning to shorten time somewhere in between (either by increasing flow rate or higher organic %) when your analytes are eluting you may observe differences. Especially with interferences. Once the method has been validated it is best not to change parameters. What is your column temperature?

We actually do this as a 3rd year lab at our university. You want to optimize your chromatogram such that all of your peaks elute with baseline separation (this may involve flow rates, and gradients). After this point, you no longer need your chromatogram, so you can delete or stop the gradient at whatever time that may be. Once the method is determined that provides the best separation, changing your gradient after this point will affect peak shape and this can affect LOD/LOQ. You're LOD I believe is calculated as  3*SD of the blank (or baseline), and LOQ i believe is 10*SD of blank. Therefore as mentioned above, changing peak shape/height can affect quantification and LOD/LOQ. Hope this helps.

I would advise carrying out a check before you proceed with this and demonstrate that there is no change. In general if you change a method you have to prove it works.  Ideally analyse 20 different samples not spiked and spiked with your analytes.  This will tell you if you have any matrix issues and will give you a good indication of your method precision. This should be compared on the old and new methods.    The time at the end of the gradient has two functions (1) remove late eluting matrix components which you probably do not see in your MRM method and (2) as Edward pointed out stabilisation time which will affect S:N.  (1) uses the a higher organic content for RP separations and (2) is your starting gradient solvent.