Many large, disease-related biobanks of serum samples have been established prior to the widespread use of proteomics in biomarker research. These biobanks may contain relevant information about the disease process, response to therapy or patient classifications especially with respect to long-term follow-up that is otherwise very difficult to obtain based on newly initiated studies, particularly in the case of slowly developing diseases. An important parameter that may influence the composition of serum but that is often not exactly known is clotting time. We therefore investigated the influence of clotting time on the protein and peptide composition of serum by label-free and stable-isotope labeling techniques. The label-free analysis of trypsin-digested serum showed that the overall pattern of LC-MS data is not affected by clotting times varying from 2 to 8h. However, univariate and multivariate statistical analyses revealed that proteins that are directly involved in blood clot formation, such as the clotting-derived fibrinopeptides, change significantly. This is most easily detected in the supernatant of acid-precipitated, immunodepleted serum. Stable-isotope labeling techniques show that truncated or phosphorylated forms of fibrinopeptides A and B increase or decrease depending on clotting time. These patterns can be easily recognized and should be taken into consideration when analyzing LC-MS data using serum sample collections of which the clotting time is not known. Next to the fibrinopeptides, leucine-rich alpha-2-glycoprotein (P02750) was shown to be consistently decreased in samples with clotting times of more than 1h. For prospective studies, we recommend to let blood clot for at least 2h at room temperature using glass tubes with a separation gel and micronized silica to accelerate blood clotting.