Uterine examinations provide an inexpensive and reliable postmortem alternative to monitor pregnancy rates in free-ranging elk (Cervus canadensis). However, this technique may be insensitive during early pregnancies (i.e., <20 d postconception), relies on proper collection of tissues, and may not be comparable to antemortem approaches used throughout the rest of the year. To circumvent some of these issues, the sensitivity and specificity of a commercially available serum pregnancy-specific protein B (PSPB) enzyme-linked immunosorbent assay (ELISA) was determined relative to uterine examination. From 2013 to 2017, paired serum samples and uteri were collected from 245 harvested free-ranging cow elk in Pennsylvania, US in November. Uteri were examined to determine whether the cow was pregnant, and, if so, gestation age was estimated based on embryo crown-rump (CR) length. The serum PSPB ELISA testing was then performed. Since harvested elk could not be retested, samples with optical densities close to the threshold for pregnancy determination (i.e., high-recheck samples) were considered as both not pregnant and pregnant, and analyses were performed separately under each scenario. Overall, the PSPB ELISA had a sensitivity of 95% (high-recheck considered pregnant) and 93% (high-recheck considered not pregnant), and a specificity of 91% (high-recheck considered pregnant) and 93% (high-recheck considered not pregnant) relative to uterine examinations. Based on CR length, gestation age was <14 to 55 d. Our results indicated the PSPB ELISA was an accurate serum-based pregnancy test for elk.