Statistical study of genotype assignment (carrier detection) in hemophilia B
1979; Elsevier BV; Volume: 15; Issue: 1-2 Linguagem: Inglês
10.1016/0049-3848(79)90053-7
ISSN1879-2472
AutoresJohn B. Graham, Paul Flyer, Robert C. Elston, Carol K. Kasper,
Tópico(s)Blood Coagulation and Thrombosis Mechanisms
ResumoAbstract Statistical studies were made of laboratory data obtained from obligatory carriers of hemophilia B and normal women belonging to reference groups assembled separately in California and North Carolina. Values for factor IX activity (IX:C) and factor IX antigen (IX:Ag) were available on 29 carriers of CRM − hemophilia B, 18 carriers of CRM + hemophilia B and 20 normal women from California, and IX:C levels were available on 24 carriers and 33 normal female North Carolinians. Ages to the nearest year were available on all of the women. A significant increase of IX:C with age was present in one sub-group of normal women from California but not in two other subgroups of normals nor any of the carriers. There was not an increase of IX:Ag with age in any group. Each variable was examined individually for its utility in carrier detection, and the utility of using both variables simultaneously was assessed by discriminant analysis, both linear and quadratic. The analyses showed that the most efficient (and cost-effective) method of classifying CRM − carriers was univariate discrimination based upon assay of IX:Ag alone. For CRM + carriers, bivariate linear discrimination utilizing both IX:C and IX:Ag assays was preferable in the particular body of data examined. Quadratic discrimination was actually more efficient in identifying CRM + carriers, but a statistical peculiarity makes this conclusion suspect.
Referência(s)