Measurement of alpha 1-antitrypsin in feces has been proposed as a method of diagnosing a protein-losing enteropathy. This approach makes use of an endogenous marker rather than radioisotopically labeled materials such as 51CrCl3 or 131albumin to measure protein clearance. The validity of using fecal alpha 1-antitrypsin measurement as a reflection of protein loss through the gastrointestinal tract has been demonstrated by several investigators. The authors report here the characterization of excreted alpha 1-antitrypsin and an evaluation of the immunochemical methods used to measure this protein. They find alpha 1-antitrypsin to be excreted both as a protease-antiprotease complex and in a form that is relatively unaltered compared with serum alpha 1-antitrypsin. The proportion of alpha 1-antitrypsin excreted as a complex was found to vary from patient to patient. Formation of the protease-antiprotease complex was found to decrease the apparent alpha 1-antitrypsin concentration when radial immunodiffusion or immunonephelometry were used. The observed bias was greater for radial immunodiffusion. When these methods were applied to a newborn population at risk for necrotizing enterocolitis, radial immunodiffusion was found to have better sensitivity and a higher predictive value for a positive result than the nephelometric method. The use of fecal alpha 1-antitrypsin for diagnosis of protein-losing enteropathy appears to be best accomplished by radial immunodiffusion.