Nearly 10 years ago, we and others suggested that the PiMZ phenotype might be a significant risk factor leading to the development of chronic obstructive lung disease. That hypothesis, based on observations that the intermediate deficiency was found more often in patient groups than expected by chance alone (1, 2), remains controversial. In a Note published in the January 1978 issue of this journal, Lebowitz and associates (3) reported that lung function test results in community volunteers with deficient Pi phenotypes were indistinguishable from the results obtained in subjects with the normal PiM phenotype. Similar results have been obtained from other population studies (4-6), suggesting that mild forms of antitrypsin deficiency are not important risk factors. Other studies, performed in patients and selected healthy volunteers, tend to support the opposite conclusion. A careful analysis of the extensive literature that now addresses itself to this question fails to resolve the controversy. However, I believe this literature does indicate why the studies reported to date have not yielded definitive answers. The PiZ phenotype is sufficiently rare so that even an occasional concurrence of the severe deficiency and emphysema left no question that this state is associated with a marked increase in a person's risk of developing chronic obstructive lung disease. Determining whether such a relationship exists for milder forms of the deficiency is a more difficult problem. The PiMZ phenotype, for example, is found in approximately 3 per cent of certain populations. All of the phenotypes that are associated with decreased concentrations of circulating antitrypsin can oci Supported by National Institutes of Health Grant HL 12833. cur in as many as 10 per cent of live births. Thus, mild forms of the deficiency could be seen as a chance occurrence in a high proportion of patients with obstructive disease even if no excess risk existed. Methodologic problems further complicated the interpretation of early studies. Electrophoretic methods have been developed and standardized; these permit reliable identification of carriers of deficient Pi types. Only studies that use such methods can help us resolve this controversy.

Standardized electrophoretic methods have now been applied to groups of patients. In the most recently published survey of this type, Shigeoka and co-workers (7) summarized the available literature. Surveys have been done in various parts of the United States as well as Germany and Scandinavia. In total, almost 1,400 patients with obstructive lung disease have been studied; 6.2 per cent were found to have the PiMZ type. This is a highly significant excess prevalence even when compared, for example, to Scandinavian populations, in which one would expect a 2.9 per cent rate of this phenotype, the highest rate found in large control populations. Although these studies suggest that the MZ phenotype increases the risk of developing obstructive lung disease, computation of a risk ratio requires a different experimental design. A 2-to 3-fold increase in the number of persons with the PiMZ type among patients with obstructive disease seems to imply that carriers of the deficiency have approximately 2 to 3 times the" normal" risk of developing this disease during their lifetime. This may not be the case. Excess prevalence of the deficiency would be observed if risks were equal in MZ and M subjects but if MZ subjects, on the average, developed overt disease or sought medical care at an earlier age. Indeed, these pa-