Simple correction improving long-term reproducibility of HPLC-MS.

Chromatographic peak areas in long series of high-performance liquid chromatography-MS experiments often vary, which decrease reproducibility and may cause bias in the results. It was found that the sensitivity of various components change differently; in our case, variability is in the order of 20-40%, and it is most likely due to changing conditions in electrospray ionization (ESI). The most often used peak area correction methods do not take this effect into account. The change in peak areas can be well described by a polynomial function; we found that a fourth-order polynomial is most often suitable. We suggest a simple correction algorithm based on polynomial fitting. When the experiments were inherently well reproducible, this correction improved reproducibility from 12% to 3% (on average for various components). When random errors were larger, this improvement was less significant (15% to 12% in nano-ESI) but nevertheless essential in order to avoid possible bias in the results.