We developed a bedside scoring system for diagnosis of trisomy 18 in the immediate neonatal period. Points are assigned for the presence of features known to occur in trisomy 18: five points for the presence of features previously reported in 50% or more of affected infants; three points for features reported to occur in between 10% and 50% of affected individuals; and one point for features known to occur in less than 10% of infants with the disorder. Using the scoring system, we evaluated two cohorts of patients: those in whom a diagnosis of trisomy 18 was previously established (retrospective group) and those in whom the diagnosis was suspected but not yet proved (prospective group). The average score in the retrospective series (n=25) was 96.7, and no patient scored less than 70. Twenty-two patients were evaluated prospectively; in all cases the presence or absence of trisomy 18 was correctly predicted. The average score in the 11 patients without trisomy 18 was 41.4, and all patients scored 60 or less. In the 11 patients confirmed to have trisomy 18, the average score was 94.3, with a range of 70 to 113. This scoring system is an accurate, reproducible method for predicting trisomy 18 in neonates with multiple congenital malformations.