The urinary lutropin (luteinizing hormone; LH) profiles of 20 patients were determined in 10 consecutive urine samples (n = 200) with a manual technique (MAIAclone, Serono; y) and a fully automated (Abbott IMx; x) technique. For 132 single determinations within the range of detection, the regression line was y = 1.39 x - 0.07 (r = 0.85). Retrospective analysis of the LH profiles obtained by the IMx yielded the classification "not pregnant" (negative) in 14 cases, of which 13 were correctly classified; 6 profiles were classified as "pregnant" (positive), which correlated with the clinical findings in 4 cases. Nine of the LH profiles analyzed with the MAIAclone method were classified as negative, which was correct in seven cases. Retrospective analysis of 11 LH profiles established with the MAIAclone assay gave positive results, which agreed with the clinical outcome in three cases. The diagnostic sensitivity of the IMx and MAIAclone methods was 87% and 47%, respectively; diagnostic specificity was 80% and 60%, respectively. In smaller series (< or = 40 samples), the IMx method has considerable advantages over the MAIAclone method in cost and convenience, because standard curves can be stored and assays can be performed as single determinations.