Due to the central role in predicting response to herceptin and possibly also other anticancer drugs, accurate and reproducible detection of the HER2 status is important. Fluorescence in situ hybridization (FISH) and immunohistochemistry (IHC) are the most commonly used methods for HER2 analysis. It is a disadvantage of FISH that a fraction of cases remain not interpretable probably due to suboptimal tissue handling before analysis. To investigate a possible influence of tissue damage on the results of HER2 IHC we compared the HER2 IHC results obtained in tumors with and without interpretable FISH in a breast cancer tissue microarray. The HER2 IHC results differed greatly between 1551 tumors with interpretable HER2 FISH signals and 405 breast cancers showing no FISH signals. FISH informative tumors had an IHC score of 3+ in 12.6%, 2+ in 3% and 1+ in 9.2% of cases. FISH non-informative tumors showed significantly lower IHC scores (p < 0.0001). They were IHC 3+ in 3.9%, 2+ in 3.7% and 1+ in 4.4% of cases. Overall, the data show that not only FISH but also IHC results are dependent on good tissue quality for successful analysis. Poor tissue quality can be easily identified in FISH analyses because of a lack of hybridization signals. Inappropriate tissue handling is more dangerous in IHC because an artificial lack of staining can be regarded as 'negative' result.