We report on the reliability of polymerase chain reaction (PCR) for the detection of Entamoeba histolytica from formalin-fixed, paraffin-embedded tissue in comparison with microscopy and have determined predictors that may influence PCR results. E. histolytica-specific and Entamoeba dispar-specific real-time PCR and microscopy from adjacent histologic sections were performed using a collection of formalin-fixed, paraffin-embedded tissue specimens obtained from patients with invasive amebiasis. Specimens had been collected during the previous 4 decades. Association of sample age, parasite density, and reliability of PCR was analyzed. E. histolytica PCR was positive in 20 of 34 biopsies (58.8%); 2 of these 20 were microscopically negative for amebae in neighboring tissue sections. PCR was negative in 9 samples with visible amebae in neighboring sections and in 5 samples without visible parasites in neighboring sections. PCR was negative in all specimens that were older than 3 decades. Low parasite counts and sample ages older than 20 years were predictors for false-negative PCR results. All samples were negative for E. dispar DNA. PCR is suitable for the detection of E. histolytica in formalin-fixed, paraffin-embedded tissue samples that are younger than 2 decades and that contain intermediate to high parasite numbers. Negative results in older samples were due to progressive degradation of DNA over time as indicated by control PCRs targeting the human 18S rRNA gene. Moreover, our findings support previous suggestions that only E. histolytica but not E. dispar is responsible for invasive amebiasis.