Factors that inhibit the amplification of nucleic acids by PCR are present with target DNAs from many sources. The inhib- itors generally act at one or more of three essential points in the reaction in the following ways: they interfere with the cell lysis necessary for extraction of DNA, they interfere by nucleic acid degradation or capture, and they inhibit polymerase ac- tivity for amplification of target DNA. Although a wide range of inhibitors is reported, the identities and modes of action of many remain unclear. These effects may have important im- plications for clinical and public health investigations, espe- cially if the investigations involve food and environmental screening. Common inhibitors include various components of body fluids and reagents encountered in clinical and forensic science (e.g., hemoglobin, urea, and heparin), food constitu- ents (e.g., organic and phenolic compounds, glycogen, fats, and Ca21), and environmental compounds (e.g., phenolic com- pounds, humic acids, and heavy metals). Other, more wide- spread inhibitors include constituents of bacterial cells, non- target DNA and contaminants, and laboratory items such as pollen, glove powder, laboratory plasticware, and cellulose. This review discusses the findings of many studies related to clinical, food, and environmental microbiology, including ap- proaches that have been used to overcome inhibition and fa- cilitate amplification for detection and typing. Few areas of biological science remain untouched by the invention of PCR (34, 81, 99). Other methods for amplifying nucleic acids (72, 123), such as Qb replicase (18), ligase chain reaction (13, 128), single-stranded sequence replication (17, 47), strand displacement amplification (126, 127), and nucleic acid sequence-based amplification (23, 122), have been de- scribed, but these methods have received less attention. Problems sometimes occur with PCR, however (124). De- spite early indications of great sensitivity, the sensitivity of PCR may be a negative aspect of the procedure, since the most commonly reported problem is false-positive results due to cross-contamination (98, 124). This problem can be overcome by UV irradiation (100), with sodium hypochlorite (92), and by photochemical or enzymic methods (25, 36, 40, 78). One problem that is less discussed is reaction inhibition. This may be total or partial and can manifest itself as complete reaction failure or as reduced sensitivity of detection. In some cases, inhibition may be the cause of false-negative reactions, since few workers incorporate internal controls in each reac- tion tube. Early evidence of exquisite sensitivity with mamma- lian cells (53) involving detection of a single molecule of DNA from a hair was not reproduced when PCR was applied to many microbial (and some mammalian) situations, where poor sensitivity, specificity, and reproducibility have been reported (16, 82, 86, 129, 132, 134). There may also be potentially important effects in PCR typing reactions (121), and difficulties can occur in post-PCR manipulation (61). Although systematic study of inhibition has seldom been the focus of published investigations, many workers have reported these effects in the course of other studies (12, 19, 21, 124, 129, 132, 133). Con- sidering the prevalence of this problem, it is surprising that few systematic and mechanistic studies of PCR inhibition have been reported. Rossen et al. (97) contributed the most com- prehensive study of PCR inhibition, identifying inhibitory fac- tors in foods, bacterial culture media, and various chemical compounds. These inhibitory factors included organic and in- organic chemicals, detergents, antibiotics, buffers, enzymes, polysaccharides, fats, and proteins. This review lists and discusses inhibitors and methods that can overcome the attenuation of amplification in clinical, food, and environmental microbiology. It is beyond the scope of this paper to discuss in detail the various physical, enzymic, and chemical methods used in the extraction, purification, and quantitation of nucleic acids. Those methods are presented and discussed in commercial literature and elsewhere (14, 95, 96, 106, 134).