Retrospective analysis of DNA from paraffin-embedded fixed bone marrow biopsy specimens is possible if preceded by amplification of the DNA sequences of interest by the polymerase chain reaction (PCR). These fixed specimens yield degraded DNA that may not be suitable for direct analysis by conventional digestion and hybridization methods. This limitation is circumvented by PCR amplification and subsequent analysis of the amplified products. The model used in this study is the amplification of a 725 base-pair (bp) beta-globin gene sequence encompassing the sickle-cell anemia point mutation, followed by Cvn I digestion. The beta A beta A, beta A beta S, and beta S beta S genotypes are derived from analysis of the allele-specific digestion patterns. Two fixatives were compared: neutral-buffered formalin and a mercury-based fixative (B-5) routinely used for bone marrow biopsies. DNA extracted from B-5-fixed bone marrow specimens was found to be more degraded than DNA from neutral-buffered, formalin-fixed bone marrow aliquots from the same specimens. PCR amplification of the 725 bp beta-globin gene sequence was successful with DNA from formalin-fixed bone marrow specimens, but not with DNA from B-5-fixed identical specimens. Analysis of the amplified product by Cvn I digestion resulted in correct genotype derivation for all patients, normal controls and positive controls (patients diagnosed with sickle-cell anemia or trait). These results indicate that intermediate-size DNA sequences can be amplified and analyzed when DNA is extracted from formalin-fixed bone marrow specimens.(ABSTRACT TRUNCATED AT 250 WORDS)