To confirm the production of IL-1 beta and to optimize detection and semiquantitation of IL-1 beta mRNA by polymerase chain reaction (PCR) techniques in skeletal muscle tissue, immunocytochemistry, immunoblotting and several procedures of RNA extraction and reverse transcription (RT)-PCR amplification were used on muscle samples from 12 patients with conditions associated with local production of IL-1 beta (AZT myopathy: 6 patients; sarcoid myopathy: 6 patients) and from 9 patients with normal muscle used as controls. Abundant IL-1 beta immunoreactivities, corresponding to both pro IL-1 beta and mature IL-1 beta as assessed by immunoblotting, were observed in all diseased muscles, either in inflammatory cells (sarcoid myopathy) or in atrophic muscle fibers (AZT myopathy). Acid guanidinium isothiocyanate phenol-chloroform extraction of RNA appeared less efficient for IL-1 beta mRNA detection by RT-PCR than proteinase K digestion followed by phenol-chloroform extraction. Even using the latter procedure, RT-single PCR for IL-1 beta mRNA was puzzlingly negative in all cases but one; in contrast, RT-nested PCR specified by DNA enzyme immunoassay yielded detection of IL-1 beta mRNA in all diseased muscles and in occasional controls, including the expected PCR product of 391 bp, but also another product of 935 bp, corresponding to IL-1 beta mRNA with unsplicing of the fourth intron. Semi-quantitative PCR showed that production of IL-1 beta mRNA was higher in sarcoid myopathy than in AZT myopathy, and in AZT myopathy than in controls. In conclusion, IL-1 beta expression can be reliably studied using immunocytochemistry, but assessment of IL-1 beta mRNA production in muscle tissue requires optimized extraction and RT-PCR procedures.