almost invariably associated with the disorder. Scientists had proven that the low-phenylalanine diet promptly lowered serum phenylalanine levels in infants with PKU. However, at the time mass screening began, there was no conclusive evidence that the diet prevented mental retardation. Because of the claims of benefit, it was impossible to mount a randomized controlled trial to determine the effectiveness of the diet. (In retrospect, 14 children detected by screening, and meeting rigorous criteria for PKU, would have been needed in a two-year trial to demonstrate the effectiveness of the diet; Holtzman, 1977.) Only after 10 years did a collaborative trial (in which all infants were treated with the special diet, but to varying degrees) provide convincing evidence of an association between early institution of the low-phenylalanine diet and maintenance of good dietary control on the one hand, and intellectual performance on the other (Committee on Genetics, 1982; N. Holtzman et al., 1986).
In the meantime, it also became apparent that not all infants with elevated phenylalanine had classical PKU. Some infants had variant forms of PKU that resulted in seizures, retardation, and death even with dietary restriction of phenylalanine; fortunately, these variants are very rare (Scriver et al., 1989). In other infants, the elevation of phenylalanine was transient or, if it persisted, was mild and not associated with mental retardation. A few infants who did not actually have PKU died or suffered irreversible damage when they were started—without proper monitoring—on the low-phenylalanine diet, resulting in depletion of their supply of phenylalanine, an essential amino acid (Holtzman, 1970).
A 1970 survey of health department and PKU clinics indicated that 5 to 10 percent of infants with PKU were being missed by screening and were being diagnosed because of mental retardation. Although it is not altogether clear what accounts for some infants being missed, two factors contribute significantly: (1) the inherent lack of sensitivity of the screening test, particularly when it is performed the first day after birth, and (2) human errors in the performance of the test, which include errors in specimen collection, laboratory procedure, or follow-up, such as failure to notify parents or the infant's physician (C. Holtzman et al., 1986). The current practice to reduce the length of stay of newborns in order to reduce hospital costs may well contribute to an increasing rate of false negatives with the PKU screening test.
Thus, screening for PKU had been undertaken widely before either the validity of the test, the quality of the laboratories performing it, or the efficacy or safety of the treatment had been firmly established. As a result, some infants suffered irreparable harm; however, many more infants benefited from early PKU detection and appropriate therapy.
A more serious set of errors was made in the establishment of sickle cell screening programs in the African-American community in the early 1970s. Some