Several reports have shown that a long delay between cutting sections and immunohistochemical (IHC) staining can decrease the IHC reaction intensity. However, systematic large-scale studies to investigate to what extent this problem may influence the outcome of translational research studies are lacking. In this study, we used a tissue microarray (TMA) approach to investigate the influence of slide age on comparisons between the results of IHC analyses for estrogen receptor (ER), progesterone receptor (PR), cyclin D1, HER2 (HercepTest), and E-cadherin and clinical outcome in a series of 522 breast cancer patients. Old TMA sections stored for 6 months at 4 degrees C and freshly cut sections were analyzed under exactly identical experimental conditions. As compared to results obtained on freshly cut sections, the frequency of positivity on old sections decreased from 65 to 46% for ER (P<0.0001), from 33 to 18.5% for PR (P<0.0001), from 16.3 to 9.6% for HER2 (P=0.0047), from 45.1 to 37.7% for cyclin D1 (P=0.10), and from 58.9 to 32.9% for E-cadherin (P<0.0001). Despite the lower fraction of positive cases, most associations between IHC data and tumor phenotype that were observed in fresh section analysis were also found when old section data were analyzed. The results confirm that slide aging has a great influence on the intensity of IHC staining in individual cases, but they also suggest that many clinicopathological associations can be detected if suboptimally processed sections are used for IHC.