In the United States, testing for syphilis traditionally has consisted of initial screening with an inexpensive nontreponemal test, then retesting reactive specimens with a more specific, and more expensive, treponemal test. When both test results are reactive, they indicate present or past infection. However, for economic reasons, some high-volume clinical laboratories have begun using automated treponemal tests, such as automated enzyme immunoassays (EIAs) or immunochemoluminescence tests, and have reversed the testing sequence: first screening with a treponemal test and then retesting reactive results with a nontreponemal test. This approach has introduced complexities in test interpretation that did not exist with the traditional sequence. Specifically, screening with a treponemal test sometimes identifies persons who are reactive to the treponemal test but nonreactive to the nontreponemal test. No formal recommendations exist regarding how such results derived from this new testing sequence should be interpreted, or how patients with such results should be managed. To begin an assessment of how clinical laboratories are addressing this concern, CDC reviewed the testing algorithms used and the test interpretations provided in four laboratories in New York City. Substantial variation was found in the testing strategies used, which might lead to confusion about appropriate patient management. A total of 3,664 (3%) of 116,822 specimens had test results (i.e., reactive treponemal test result and nonreactive nontreponemal test result) that would not have been identified by the traditional testing algorithms, which end testing if the nontreponemal test result is nonreactive. If they have not been previously treated, patients with reactive results from treponemal tests and nonreactive results from nontreponemal tests should be treated for late latent syphilis.